Report No: 61631-ET Document of the World Bank OPTIONS FOR STRENGTHENING LAND ADMINISTRATION FEDERAL DEMOCRATIC REPUBLIC OF ETHIOPIA (Executive Summary) (High tenure security) - (encroachment) (access to land) (land-use plan) - (vulnerable groups) (land tenure regularization) - 1 (Nationwide Sustainable Land Administration and Use System) ( 2009) (Land Admnistration and Land Use Development Project (LALUDEP) (Concept Note) ( ) ( 2010) (Urbanization) (tenure security) - ( ) ( ) ( ) . (Legal Framework) - “ � ( ) (transactions) ( ) :: - (permit) 99 2 (Negotiated deals) (prime location) (permit) (ambiguities) . . . ( ) ( ) 3 (permit) (urban-rural interface) (Structural Development Plan) ( ) ( “ � (land development fee) (land dedication for public purpose) - “ � (public purpose) “ � - (Scope) 4 (intra-city redevelopment) ( ) (procedural changes) (expropriated lands) ( 100-250 ) (expropriation) (land allocation) ( ) ( ) ( (land re-adjustment and land pooling) - ( ) (property valuation) 5 - (legal shortcomings) - ( ) “ � (legal status) ( ) ( ) “ � (land abandonment) - ( ) (sustainable land use) (land-use plan) - (large scale informality) 6 ( ) . 1 2 ( ) ( ) - ( ) ( ) ( ) ( ) 1 2 7 ( ) - . . (cadastral-index map) (aerial photogrammetry) (remotely sensed images) . . (EMA) 3 (geodetic control) (large scale) (cadastral-index maps) - (geographic information system) - 3 . www.surdex.com/MappingGlossary.aspx - http://www.fgdc.gov/standards/projects/FGDC-standards- projects/framework-data-standard/GI_FrameworkDataStandard_Part4_GeodeticControl.pdf 8 (expropriation) - ( ) (obsolete) ( ) ( ) ( ) - (feedback) - 9 (task force) ( ) - ( ) / ( ) ( ) ( ) 10 / ( ) 25 - ( ) 3.4 ( ) - ( ) - ( ) ( ) - ( ) - ( ) / 11 (protocol) (Paper based) (IT) (ICT ) (protocol to access) (data) (back up file) (software) - (target groups) (land-use planning) (realistic) - spatial) (spatial) 40 21 12 (spatially) - (record) ( ) (high- resolution satellite imagery) (spatial data) (minimum) (cadastral-index map) (orthophoto maps) (aerial photography) ( ) ( ) (large scale) (orthorectified imagery) 13 - - (multiplicity) / - (informality) - - - 14 (Matrix) • ( ) ( ) ( ) ( ) ( ) • (Informal land holdings) ( ( ) ( ) • ( ) • - • • • (land use planning ) (Guideline) ( ) ( ) • 15 • • (interoperable) (Data) ( ( ) • “ � • - () ( ) • ( ) (land abondnment ) • - (large scale informality) ( ) • • 16 • (Sptial record) ( (C.I.M) ( (Aerial Orthophoto Maps) • / • ( ) • • • • ( ) 17 • • / ( ) • / • (Technical Working Group) • / • - 18 • ( - ) (Systematic Registration) • ( ) / • ( ) • / • / / 19 • • (High Resolution Satellite Imagery) (Cadastral Index Map) / • (Orthorectified) (Cadastral Index Map) • ( ) (Data Access Protocol) (maintainance)/ (Updating) • ( ) (Communication Strategy) ( ) • 20 • (Updating Process) (Back up Copy) - • (conclusive) • ( ) • 21 EXECUTIVE SUMMARY Over the coming decades, land policy and administration, for urban as well as rural areas, will be critical for Ethiopia’s development. The vast majority of people making up the Federal Democratic Republic of Ethiopia’s (FDRE) predominantly agricultural economy live in rural areas. There, land continues to be a key household asset and functions as a safety net. High tenure security will be critical to provide the incentive to invest in land. Tenure security will increase productivity and encourage transfer of land to its most effective use, thus contributing to the growth and transformation of the agricultural as well as the transition to more urbanization and industrialization. Secure tenure also will reduce encroachment and unsustainable resource use that are leading to erosion and permanent loss of soil fertility. Access to land is also one of the most important factors for industrial growth. In addition, the ability of local, particularly urban, governments to pursue land-use planning including for promoting urbanization and industrialization and to draw on land to generate public revenue are pillars of administrative decentralization that are closely linked to transparency and good governance. Finally, land policies and administration can contribute significantly to the objectives of promoting gender equality and protecting vulnerable groups in Ethiopia. International experience shows the significant positive impact of strengthening land administration and management systems and the feasibility of doing so within the context of poor countries. Countries such as China and Vietnam demonstrate that gradual reform of the land system is possible within the system of state ownership of land. Moreover, such reforms have been shown to yield very high economic and social returns. Other examples suggest the possibility of affordable land administration reform by using technology that is appropriate to the African context. For instance, Rwanda is rolling out a nation-wide program of land tenure regularization showing already positive impacts on investment, gender, and the incidence of conflict. Now is a good time for Ethiopia to address its pressing land policy and administration issues. The country already has had significant relevant positive experience. The rural land registration program implemented in Ethiopia from the start of the last decade is one of the world’s largest. The registration is implemented equitably and with clear positive impacts on conflict, productivity, investment, and rental market participation. Based on this experience, plans for a national sustainable land administration and use system have been discussed at a high political level (MoARD 2009). A directorate for rural land administration and use under the Ministry of Agriculture (MoA) (formerly MoARD) has been established. Additionally, a concept note proposing a Land Administration and Land Use Development Project (LALUDEP) was discussed internally and with support from development partners. Finally, the high growth of the economy and the ambitious targets of the recent (draft) Growth and Transformation Plan (MoFED 2010) is increasing demand for land administration services and more efficient land management. To be sustainable, high agricultural growth and rapid urban development need to be supported through tenure security, good planning and governance, and other elements related to a strong land administration and management system. This report aims to assist the government of Ethiopia in improving the performance of its land administration system based on a detailed review and lessons from Ethiopia and other countries. The three key areas for improvements and options for improving the land administration system 1 summarized in the action matrix below are the (a) legal and regulatory framework; (b) administrative capacity and organizational set-up; and, based on these, (c) provision of efficient, cost-effective, and sustainable land administration services and land management. A. Legal Framework Ethiopia’s constitution states that ownership of all land is “vested in the State and in the peoples of Ethiopia�. Citizens receive permanent or long-term rights to access and use land. Specific laws clarify land tenure rights, regulate transactions, and establish procedures to take land under eminent domain and compensate for such takings. However, the remaining gaps in regulations and in the guidelines for their implementation (and, in some cases, their failure to harmonize with the regulations) cause variations across the country, some lack of clarity, and, hence reduced tenure security and governance. Action 1: Strengthen urban land legislation and regulation and harmonize the linkages with rural land legislation. Urban land rights are held primarily in the form of unlimited-horizon “permits� from urban authorities, which require annual “rent� payment, administratively set at low levels. More recently, land rights have been awarded under a “lease� system with a finite horizon of up to 99 years. Leases can be priced at market value by either allocation on auctions or through market-based valuation for “negotiated� deals. However, most leases have been allocated at low administrative prices for residential or social uses or at below-market levels through “negotiated� allocation for commercial uses, even if in prime locations. Moreover, ambiguities arise for buyers of properties on the leased land when they try to obtain a lease contract on their name. Consequently, leased land is generally perceived to carry less tenure security compared to permit land. Moreover, local governments and municipalities are in charge of urban land registration with little central guidance, leading variations across the country. All of these ambiguities have led to highly variable arrangements for urban land administration, particularly, the types of rights and encumbrances to be recorded, mechanisms for doing so, spatial and textual data requirements, and data management. Combined with the lack of capacity and resources, these ambiguities imply that, in many cities, land registration is incomplete and out of data, informality is high, arrangements to monitor leases and lease revenue non-harmonized and inadequate, and land management often ineffective. The potential for land revenues is hardly used. To address these factors, it is critical to finalize the drafting and ensure the passage of proclamations on (a) urban property registration, including clarification of the tenure and rights of urban residents in properties subject to secondary transactions; (b) the establishment of real property registration organs; and (c) the regulatory framework for cadastral surveyors. Passage and broad dissemination (including a request for feedback) of this legislation could help expedite implementation, initially on a pilot basis, of procedures to formalize urban informal landholdings. To start addressing the challenges of informality, these procedures should be consistent with the existing (or suitably modified) regulations. Having a functioning land registry and a clear understanding of the legal status of different types of land is also provides the government with a basis for policies on land-use planning rules and for prudent land management that recognizes existing rural land rights in processes of urban expansion. Departing from the current standstill of the dual land tenure system in urban areas toward a unified and more efficient system requires a frank discussions and decisions including on 2 respective corrections to existing proclamations and regulations. A focus should be on systematic simplification and unification of the leasing system, making it more attractive to land tenants, and on creating incentives for permit holders to switch to leases. At urban-rural interface, a fundamental problem is that there is no mechanism of converting land rights of rural landholders into urban rights: as soon as any rural territory is slotted for urbanization by inclusion into a Structural Development Plan (formerly, Master Plan) of a nearby city or urban center, land holders at this territory are assumed to be subject to expropriation. By international standards, this is quite unusual and not in the realm of good practices. Indeed, good international practices allow holders of the rural land converted into land for development to benefit from increased value of their land, but this gain is shared with the community or government in one form or another. In Ethiopia, there is an acute need in establishing a clear legal mechanism that would automatically convert legitimate land holding rights of rural tenants into urban land rights when a rural territory is planned for an urban expansion. Another critical task is to introduce rules that would allow legal holders of land in these areas to benefit from the conversion and resulting increases of land values (for example, by participating in land redevelopment or gaining from it financially). However, these new arrangements should also introduce fiscal instruments (“land development fee,� land dedication for public purposes, etc.) that would capture some part of the increased land values into a public budget and earmark this revenue for investment in public infrastructure. Action 2: Close gaps in rules and regulations as well as institutions for eminent domain and compensation. Unclear guidance on what determines a valid “public� purpose for eminent domain expropriation together with institutional gaps implies that in practice expropriation is often seen as arbitrary and inconsistent across the country. In particular, there is substantial scope for officials to define “public interest� purposes for land expropriation. Moreover, systematic land expropriation and eviction of the rural population in the areas of urban expansion has been the instrument of making land available for the urban growth. Appeals procedures, even if adhered to, often are biased against landholders and their effectiveness is further undermined by both the government officials’ and citizens’ limited awareness of rules and a paucity of independent land valuation expertise. Thus, land loss through expropriation with compensation at less than market values or without ensuring that livelihoods are restored remains a source of tenure insecurity, particularly in areas of inner-city “redevelopment� or rural areas near cities. To better understand the issue and identify possible remedies, it is important to, first, document and assess the extent to which sample expropriation cases (in rural as well as urban areas) are in line with international best practice, particularly the requirement that affected households will be able to maintain their living standard. Second, it is important to use this evidence to recommend procedural changes where needed, including to assist regions in harmonizing regulations on land valuation and structuring the compensation and for establishing administrative appeals procedures that are independent of the expropriating agency and represent all stakeholders. Then, based on realistic estimates of administrative capacity needed, the government should put in place programs to build required expertise. Moreover, most of expropriated urban lands have been allocated for urban housing at low administrative prices to applicants from ever-growing waitlists, according to relatively high land consumption rules (100 – 250 m2 per household). This practice is not sustainable fiscally, spatially, and socially. Therefore, reconsidering both expropriation practices and land allocation 3 practices should be a priority. Reduction of land expropriation to an avoidable minimum (for public infrastructure) should become a policy and practice in areas of the urban expansion. This can be achieved through a combination of tools, including through better targeted urban planning (selection of urban growth areas where minimal expropriation will be needed) and use of voluntary participation techniques (“land re-adjustment� and “land pooling�). Finally, to improve compensation for expropriation, reduce unfair gaps in land payments in urban areas, and increase revenues from urban land holders (both “permit� holder and lessees), building an effective, transparent market-based system of land pricing based is desirable. This system would be based on gradual approximation of market values and has to include capacity building for implementation. In particular, this should include training on property valuation: basic for government officials and more advanced for the private sector real estate professionals. Action 3: Address the remaining limitations to rural land legislation and regulation, especially in “emerging� regions. The legislation of rural land rights at the federal level and in regions such as Amhara, Oromiya, SNNPR, and Tigray – and its implementation, particularly through the certification and registration of rural land rights – has had major positive impact. The system could be further strengthened by addressing some remaining legal shortcomings. These comprise (a) the ambiguous status and unclear definition of communal lands; (b) limits on land leasing that vary across regions; (c) a wide and sometimes ambiguous definition of land abandonment; and (d) unrealistic minimum holding sizes. These legal gaps and limitations undermine sustainable land use, encourage informality and encroachment, and reduce tenure security. Shortcomings vary by region and exist particularly in the “emerging� regions. Afar, Gambella, and Benishangul-Gumuz regions have developed and adopted (rural) land use and administration proclamations but have not yet commenced implementation. Ethiopia’s Somali region has not as yet been able to develop or adopt proclamations. The ongoing work to overcome these gaps should continue. The adoption of legislation for pastoral and agropastoral areas is particularly challenging and requires careful study and a participatory process. Such legislation needs to clarify the legal status of customary systems and recognize the communal nature of associated resources. This legislation should be combined with the definition and establishment of institutional arrangements. They could include the definition and registration of user groups and authorized representatives of such groups, model bylaws outlining mechanisms for groups to discipline members and ensure adherence to agreements, and ways to individualize land tenure if the group consensus exists to do so. If such arrangements are in place, it will be possible to use results from systematic land-use planning to record rights to different types of land and to ensure that these rights are recognized. Doing so will be particularly important if land is required for other purposes, including to deal with demand for land from outside investors. Evidence in Ethiopia and other countries suggests that the restrictions imposed by some regions impede efficiency-enhancing land rental market transactions. It therefore is desirable to explore options to gradually eliminate existing restrictions on land leasing. It also is important to clarify that temporary land leasing does not imply abandonment and thus should not result in land being taken back. Limitations on holding sizes should be defined, if at all, in a way that ensures that they do not lead to large-scale informality and that they do explore options (including incentives) for voluntary land consolidation. 4 B. Organizational Set-up and Implementation Capacity Historically, weak or nonexistent central institutions, wide variation in capacity across regions, and a virtually complete institutional separation between rural and urban institutions have made effective policy implementation and service delivery very difficult. The Government of Ethiopia has recognized the importance of this issue and has taken decisive steps to overcome it. At the federal level, a new directorate for land administration has been established within the Ministry of Agriculture (MoA) to address issues of rural land policy and administration.1 The Department of Land Development and Administration in the Ministry of Urban Development and Construction (MoUDC) is responsible for urban land matters (policy, planning, capacity building, and guidelines).2 Regional institutions handle the lion’s share of the day-to-day work on land administration. These institutions largely mirror the federal arrangements, including the separation of rural and urban land administration. Dealing with land administration at the local level is the responsibility of the cities, which have substantial discretion, woreda (district) and towns. Rural land matters are handled by woreda and kebele (village) staff. They report to local government and receive technical guidance from the regional government departments. In sum, the institutional set-up for land administration in Ethiopia is complex, varying by region and divided between a rural and an urban sector. Action 4: Harmonize standards and clarify responsibilities for land registration and information of different administrative levels (regions, woredas, kebeles) through increased central guidance and coordination efforts. The potential benefits of applying land administration information to a wider range of uses (for example, land-use planning and infrastructure provision) are not materialized. Moreover, the cost of maintaining this information is increased by limited data-sharing capability and incompatible technological standards. Unclear definition of responsibilities for different administrative levels or a mismatch between responsibility and capacity increases the cost and reduces the quality of providing land administration services. Lack of clear definition of responsibilities also can create overlap among administrative levels (regions, woredas, and kebeles). In addition, lack of a clear regulatory framework for private sector engagement makes it difficult to attract private sector (including paralegals and para-surveyors) participation to complement public institutions and speed up roll- out and capacity building. Clarifying responsibilities requires national institutions to take the lead. They also should lead in setting standards (including for software) that the regions might adapt to their requirements, in building capacity for their adoption and for monitoring performance. There is ample scope for improving coordination across levels of government. Improving coordination would include a. Identifying key functions to be performed by the public sector with respect to first-time registration and maintenance. b. Setting standards for spatial data that are cost effective and that enable a rapid scaling-up of land registration in rural as well as urban areas. These standards most likely would be based on cadastral index maps that, in turn, are based on aerial or remotely sensed images. 1 Until recently, the Ministry of Agriculture and Rural Development (MoARD). 2 Until recently, the Ministry of Works and Urban Development (MoWUD). 5 c. Identifying delivery models at the kebele, woreda, and municipal level that can be used by regions to perform land administration services effectively at different levels of demand and to quantify associated human and physical capacity needs. d. Designing a system to regularly monitor and publicize performance measures by land administration institutions and explore ways to link performance to eligibility for specific types of support as well as budgetary allocations. Specialized technical institutions such as the EMA will need to provide geodetic control3 and consistent large-scale maps to the regions in a timely way and based on demand. These tools will help the implementing agencies establish a low-cost spatial record (cadastral index maps, or CIMs) in rural and urban areas, which is necessary for the agencies to perform their functions. EMA also can help supervising private sector efforts. Regions initially will develop survey capacity and GIS systems and provide hard copies of maps to the woredas and kebeles. Over time, as capacity is built, the survey capacity and GIS systems may be devolved to woreda level. The general rural land administration department/unit within MoA and at the regional level should be responsible for registration of land for large-scale agricultural investment areas. In contrast, institutionally, other administrative processes of large-scale lands including expropriations clearly should be separate. Finally, responsibilities of local entities under MoA that handle rural land certification could also be expended to include systematic registration of land rights in small urban centers in rural woredas. Action 5: Institutionalize functioning arrangements for record maintenance, that is, provide clear responsibilities and guidelines, and capacitate relevant agencies. Even in cases in which the “first-level� certification––the broadly applied process of certification without much spatial reference––has been completed, the information generated easily becomes obsolete as arrangements for maintenance are unclear, too costly, or entirely lacking. To ensure that records remain current, the Government must identify delivery systems and build the associated capacity required to enable timely response to user demand, the willingness and capacity of the user to pay for services and the public-good content of different activities and associated sharing of responsibilities between the public and private sectors. Ways to support the establishment of such a delivery system in a way that ensures its long-term sustainability (including cost recovery) provide one argument for strengthening the integration of rural and urban land administrations. Doing so would ensure that fees received in high-transaction areas (especially urban) can cross- subsidize those with less activity (rural areas). The capacity of educational institutions will need to be strengthened to be able to support the anticipated human resource needs of the land sector. Also needed are advisory committees that represent all stakeholders to provide user feedback and oversight. Action 6: Strengthen rural-urban integration of land registration, information and other services, initially through coordinating mechanisms at various levels and definition of common standards. The institutional separation and lack of coordination between urban and rural land makes quick adoption of best practice difficult, increases the cost of land 3 Geodetic control means to locate map features correctly in relation to their actual locations on the earth’s surface. Geodetic control also is defined as a network of carefully measured horizontal and vertical points. An accurate geodetic network provides the foundation for photogrammetric mapping. www.surdex.com/MappingGlossary.aspx See four additional definitions at http://www.fgdc.gov/standards/projects/FGDC-standards-projects/framework-data- standard/GI_FrameworkDataStandard_Part4_GeodeticControl.pdf 6 administration, and limits the scope for land management and use, particularly in periurban areas if city expansion uses land traditionally used by farmers. Globally, many countries with similar separate structures for urban and rural lands faced similar problems and opted for a more unified structure to overcome these problems. A lack of common standards between rural and urban areas leads to a fragmented and high-cost system. Their lack also leads to the neglect of rural land records during urbanization––a serious problem that undermines tenure security and fosters conflict due to nonrecognition of rural land rights. A unified rural and urban regional land administration entity should be the medium-term objective for regional governments. In the short term, it will be important that land administration be handled by specialized offices that report directly to the regional governments and are budgeted independently of sectoral offices. Arrangements for better coordination between urban and rural administrations in the regions then should be worked out, with particular attention to the handling of periurban areas. Based on international experiences in land management, it is recommended to establish a cross-agency task force or committee – at both, federal and regional level – with a clear mandate to develop realistic solutions for specific problems. These groups should be tasked with outlining a concept and implementation procedure for addressing the problems on urban-rural fringes. After their suggestions are discussed with all key stakeholders include the public and tested in a small number of localities, they can be codified in new regulation for broad implementation. Action 7: Strengthen dispute resolution institutions with a focus on building the capacity of informal/traditional mechanisms and clarifying their relationship to the formal system. Informal traditional arrangements to resolve conflict are recognized by law and studies suggest that the public, particularly in rural areas, prefers mediation of disputes through traditional local procedures. In order to use these mechanisms more widely and effectively in reducing the frequency of conflict, the individuals involved need to be trained in both the local rules and the applicable formal legislation and have access to relevant information including key pieces of legislation. As the frequency and incidence of conflicts can help illuminate areas in which the law is not well understood, it would be useful to conduct a study (or establish a typology) of the conflicts handled by different institutions, quantify and identify the most frequent types, and assess the adequacy of current (formal as well as informal) institutions to deal with them. In some cases (for example, public land encroachment), the frequency of conflicts can be cross- checked against the incidence of objective violations as obtained, for example, from satellite imagery at different points in time. This study also can help to identify areas in which there may be the need to clarify legislation; issue regulations; or educate members of traditional bodies, the judiciary, or the public at large to reduce the frequency of conflicts. Finally, the study can help to elaborate a strategy to build capacity in the judicial system and develop material to educate members of informal conflict resolution bodies about the applicable legislation and local rules. C. Providing Effective Land Administration Services As part of an effort to enhance rural tenure security, the four main regions (Amhara, Oromia, SNNPR, and Tigray) undertook a program that registered approximately 25 million parcels in kebele land books, and provided landholders with certificates that provide a textual record and identify parcels boundaries by listing adjoining neighbors. This program was carried out in a decentralized, participatory, equitable, and transparent manner building on an elected Land Administration Committee (LAC) with broad geographic participation. The program thus was 7 able to handle the massive numbers of registered holdings at low cost and within a short timeframe. Studies show that the program increased farmers’ perceived level of tenure security, investment incentive, and level of land market participation while reducing conflict and helping to empower women through a number of channels. The Government should build on this success. Action 8: Establish arrangement to ensure regular and sustainable maintenance/updating of land records. Arrangements to maintain records are lacking in many of the instances in which it has been completed. However, establishing records without the capacity to maintain them is not sustainable. In Tigray, for example, many of the 3.4 million records generated are judged to have by now become obsolete. Dealing with these issues in a realistic timeframe will require a major effort that is comparable in magnitude to the first phase of land certification. Thus, the capacity to maintain records should be the key for the pace of certification and any strategy to expand coverage with first- or second-level certificates should be based on demonstrated capacity for maintenance. Key areas of concern relate to the (a) format of the registry book and data structures and procedures for updating; (b) staffing, assignment of responsibilities at different levels, and matching the requirements to capacity; and (c) limited private sector involvement. To create the preconditions for low-cost maintenance, it will be necessary to establish a format for registry books and other documents to enable easy updating (including essential backups) in a way that can accommodate the different regional arrangements. In revising the data models used in rural areas, it also will be important to move toward establishing a common standard that can accommodate (with slight adaptations if needed) different practices in use at the regional level. Based on the above, establish protocols for sharing, processing, and maintenance of textual as well as spatial data at kebele, woreda, municipality, and regional levels; and assess the implications in terms of staffing and institutional structure. To the extent that an improved paper- based system will need to form the basic document layer at the woreda and kebele level, the human capacity to handle it will need to be made available and storage facilities to be upgraded as needed. Even if a roll-out of IT is unlikely in the near future, a well-thought out ICT strategy can help define data structure; protocols for data access, integrity, and interoperability; and progress monitoring to integrate existing models and structures and provide data back-ups. Once relevant software has been developed, it can be implemented at the regional level and rolled out from there. This strategy can help clarify the process for updating registry books, for providing information to the public and developing communication strategy for different target groups. A second issue to be resolved is to ensure that the need for record maintenance correlates with the staffing at different levels. Virtually all regional offices have significant gaps between actual and required staff numbers at the headquarters and woreda levels. Moreover, with existing staff often being trained in land-use planning rather than in land administration, the skills gaps may be large. For staff to perform these transaction services effectively, a large training is needed. Training would entail specialized in-house courses, short-term diploma training in appropriate institutes, as well as long-term and specialized university degree training in the technical, administrative, and managerial aspects of land administration. Given the massive demand, coordination with and among universities and other training institutions will be required. As reliance on the public sector alone for data maintenance is unrealistic, it will be important to also come up with ways in which the private sector can be drawn in. The government, in close consultation with the private sector, should put in place such mechanisms that build private sector capacity, that establishes data access protocols, and that ensure effective public oversight 8 including over accreditation, fees, and charges to be levied by the private sector as well as the validity of data generated by the private sector. Action 9: Complete rural land records and add a spatial element in a phased approach. More than 40 percent, that is, an estimated 21 million plots, have not yet undergone first-level certification. These likely are to be located in areas in which such a program is more difficult to implement or capacity is deficient. The fact that rural land records lack a spatial element makes it difficult to use land information to identify and protect communal areas from encroachment, to effectively plan the use of land, and to ensure completeness as well as updating. Such an effort will be warranted only if there is agreement on an approach that is sufficiently fast and low cost, can create records that can be maintained and updated locally with existing capacity, is intuitive and can gather widespread acceptance, and can be integrated with urban land records. A number of African countries (for example, Rwanda for individual or Namibia for communal rights) developed systems based on high-resolution satellite imagery that meet these requirements. Before launching into a potentially very costly effort to create second-level certificates, it will be essential to agree on the methodology for generating spatial data and an institutional model to maintain and update them; and to identify the cost of first-time registration and minimum staffing and technical requirements for woredas and kebeles to be able to maintain records. Based on international experience, the preferred method to do so will be cadastral index maps based on orthophoto maps produced from either high-resolution satellite imagery or aerial photography. Urban areas already have decided to use large-scale orthophoto maps from aerial photography as a spatial reference. There is, however, a need in the rural areas to define a process for mapping and adjudication at a large scale that builds on the strengths of the first-level certification process, systematically covers communal areas, and could be linked to simple land- use planning. Once this process has been defined, defining criteria in terms of demand for spatial information (or the need to update or improve on first-level information) could guide adoption of this process systematically. It may be worth considering applying second-level certification directly, even in areas in which first-level certification has not been completed and that are assessed as high priority. Experience and procedures with use of large-scale orthorectified imagery from other countries (for example, Rwanda, Thailand) should be drawn on systematically. It will be necessary to identify the reasons why certain kebeles have not been covered by first-level certification and use this information to inform strategies to complete first- level certification, possibly using a modified methodology incorporating spatial information based on priority in a way that links to maintenance requirements. Although global experience suggests that demand for more precise survey on a full-cost recovery basis is quite limited, it will be important to define methodologies to meet such demand, ideally through the private sector. Action 10: Provide clear and transparent ownership records in urban areas, thereby reducing informality. Urban areas are characterized by either a multiplicity of land-related documents with unclear legal standing and/or high levels of informality with little prospect for formalization. Unlike the large-scale rural program, the government has made no such systematic effort to register urban properties. The likely most ambitious program to overhaul the land administration system is currently implemented by Addis Ababa. However, most initiatives are taken up ad hoc in some cities or municipalities and in many cases have not been completed. Given the high level of informal transactions in urban areas, and the accelerating pace of urban migration and population growth, this situation requires the government to take remedial action through cadastral mapping and regularization. In addition, it would be useful to analyze past and 9 ongoing efforts to establish cadastre systems and regularize urban land rights in different contexts. Regularization should include an assessment of the potential to integrate different land- related documents to establish land-use records. There are positive examples in Ethiopia and internationally to draw from. The example of formalizing informal tenants in Mekele should be studied and disseminated, to help establish regulations and practices of effective, efficient, and humane approaches to integrating informal settlers in the formal sector. All of these efforts should facilitate arriving at operational models that could be used to design mechanisms to establish and manage such records on a large scale and in ways that would build on existing information and records as much as possible. 10 Action Matrix to Strengthen Land Administration Part I. Legal and Regulatory Issues Action Responsibility (lead) Priority Timing Action 1: Strengthen urban land legislation and harmonize the linkages with rural land legislation. • Finalize drafting proclamations on (a) urban property registration, (b) establishment of real MoUDC property registration institutions, and (c) regulatory board for cadastral surveyors (in urban High ST (then regions) areas); and move toward implementation. • Define procedures to formalize or legalize informal landholdings in urban areas in which MoUDC Medium MT occupation is consistent with existing (or suitably modified) rules and regulations. (then regions) • Discuss options for integrating the current dual land tenure system in urban areas in a single Cities, regions, High MT system feasible to implement and operate. federal level • Design conceptually and then legally a mechanism of automatic conversion of legal land rights MoUDC and MoA of rural landholders into urban rights when the rural territory is slotted for urbanization by High MT jointly being covered by urban planning documents • Design and test prototype mechanisms and fiscal instruments for landholders in these areas and local governments to benefit financially from increased land value in areas converted from MoUDC Medium MT rural to urban • Establish procedures to formalize informal development in peri-urban areas MoUDC and MoA Medium MT • Prepare guidelines on application of the Urban Planning Proclamation on land-use planning MoUDC, with (reflecting aspects comprehensively, including social and environmental and including focus contributions from Medium MT on periurban areas; recognition of existing land rights and minimize need for expropriation. cities, regions Action 2: Close gaps in legislation as well as institutions for eminent domain and compensation, including valuation; improve pricing of urban land and revenue management; introduce more sustainable approaches to land expropriation and allocation. • Based on an assessment of a sample of expropriation cases (in rural as well as urban areas) recommend procedural changes to ensure that compensation enables maintaining living MoUDC, MoA High ST standard. • Assist regions in developing regulations and directives and ensuring that consistent valuation MoUDC, MoA High MT and compensation regulations are applied to expropriation cases. Regions (with • Create interoperable data standards and ensure recognition of rural land rights in case of MoUDC/MoA Medium MT urbanization or other land takings. guidance/support) 11 • Establish a market-based system for land valuation in urban areas and an effective and transparent land valuation system based on use value in rural areas; build the capacity to MoUDC High MT broadly implement these systems. Action 3: Close the gaps in rural land legislation and regulation, including for communal lands, and for emerging regions. • Conduct an assessment of the extent of encroachment and overuse on pastoral and other communal lands to identify the underlying factors. Use this data as well as focus groups in different regions to identify ways to increase incentives for sustainable management (for MoA, MoFA, regions High ST example, model bylaws for user groups, individualization, and land-use plans to identify current uses). • Based on a review of evidence in Ethiopia and elsewhere, explore options to gradually eliminate restrictions on land leasing and to clarify that land leasing does not imply land MoA Medium MT abandonment. • Define limits on holding sizes, if any, to ensure that they do not lead to large-scale informality; MoA Medium MT explore options (including incentives) for voluntary land consolidation. • In emerging regions, based on careful study of relevant local arrangements for land administration and users’ demands, put in place a participatory process to draft regional Regions, with support High ST proclamations, regulations, and directives that are in line with the federal proclamation and the from MoA, MoUDC social and political contexts of the relevant regions. • Identify and strive to eliminate inconsistencies between federal and regional proclamations; MoA, MoUDC, Medium ST/MT and across regions. regions 12 Part II: Organizational Set-up and Administrative Capacity Action Responsibility (lead) Priority Timing Action 4: Harmonize standards and roles and responsibilities of different administrative levels (regions, woredas, kebeles) through increased central guidance and coordination efforts. • Through Ethiopian Mapping Authority (EMA), provide geodetic control and consistent large- scale maps in a timely way to regions based on demand to help them establish a consistent EMA High MT low-cost spatial record (cadastral index map, or CIM) in rural and urban (aerial orthophoto maps) areas. • Based on experience and/or pilots, identify the key functions to be performed by the public MoA High ST sector with respect to first-phase registration as well as in the maintenance of land records. • Identify a set of delivery models at kebele and woreda levels that can be used by different regions to perform these services effectively. Identify the associated needs for human and MoA (with regions) High ST physical capacity. • Design a system to regularly monitor and publicize measures of performance by different land administration institutions and explore ways of linking the results to eligibility for specific MoA Medium ST types of support as well as budgetary allocations. MoA, MoUDC, • Identify areas in which the private sector can effectively complement public sector efforts. Medium MT EMA Action 5: Institutionalize functioning arrangements for record maintenance, that is, provide clear responsibilities and guideline, and capacitate relevant agencies. • Based on user demand and capacity to pay, as well as the public good content of different MoA, MoUDC High ST activities, identify overall capacity needs (in private or public sector). • Strengthen educational institutions to satisfy anticipated human resource requirements in the MoE, MoFED, MoA, High MT land sector. MoUDC • Based on the anticipated workload, identify options for the type, structure, and staffing as well as other resource needs by land offices at regional, woreda, and kebele level; explore ways of MoA Medium ST supporting their establishment (including cost recovery) to ensure sustainability. • Explore options for cost recovery in the long term in ways that does not unduly compromise MoA, MoUDC, Medium MT access or effectiveness of service delivery, including through stronger rural-urban integration. MoFED 13 Action 6: Strengthen rural-urban integration through coordinating mechanisms at various levels and definition of common standards. • Establish a coordination mechanism, for example, initially through a technical working group (TWG), for continued interaction between rural and urban land administrations at federal or MoA, MoUDC High ST regional level with representation from regions and municipalities. • Identify user needs and pilot arrangements to manage land information and transfer, MoA, MoUDC Medium MT integration of records, and dispute resolution in the expansion zone of urban centers. • Establish procedures and processes to support the establishment and operation of an integrated MoA, MoUDC, High MT land administration agency in Oromia; evaluate experience and disseminate results. Oromiya • Task the rural registration agencies with systematic registration of land rights in urban centers MoA, MoUDC Medium MT (towns, subtowns) that administratively are subordinated to rural woredas. • Integrate registration of land for large-scale agricultural investment areas into general rural MoA, regions Medium MT land administration, while keeping separate other aspects of large-scale land administration. Action 7: Strengthen dispute resolution institutions with a focus on building the capacity of informal/traditional mechanisms and clarifying their relationship to the formal system. • Based on study (and a typology) of number and type of conflicts handled by different MoA, MoUDC, MoJ institutions, identify the most frequent types of conflicts and assess the adequacy of current Medium MT and regions (formal as well as informal) institutions to deal with them. • Use the conflict typology to identify critical legal issues in which there may be need to clarify MoA, MoUDC, MoJ legislation; issue regulations; or educate members of traditional bodies, judiciary, or public to Medium MT and regions reduce the frequency of conflicts. • Translate the above into an effective strategy to build capacity in the judicial system and MoA, MoUDC, MoJ develop materials to educate members of informal conflict resolution bodies about the Medium MT and regions applicable legislation and local rules. 14 Part III. Ensuring Effective Service Delivery Action Responsibility (lead) Priority Timing Action 8: Complete rural land records and add a spatial element in a phased approach. • Based on understanding why certain kebeles have not been covered by first-level certification complete first-level certification, possibly using a modified methodology that incorporates MoA, regions Medium MT spatial information on a priority basis. • Establish minimum standards for cadastral index maps based on orthophoto maps using high- MoA, EMA High ST resolution satellite imagery or aerial photography as the preferred option in rural areas. • In areas in which first-level certification has not been completed as well as those in which there is demand for spatial information, adopt a participatory process that establishes cadastral MoA, regions Medium MT index maps based on large-scale orthorectified photo maps. • Establish data-access protocols and regulations to check validity of data generated by private MoA, MoUDC, Medium MT sector players such as private surveyors. EMA Action 9: Establish arrangement to ensure regular and sustainable maintenance/updating of land records. • Design a communication strategy that encompasses different target groups (LAC members, vulnerable groups, communal landholders, women, woreda officials) based on an assessment MoA, regions Medium ST of key knowledge gaps and interaction with users (for example, through conferences as practiced in Amhara). • Establish a format for registry books and other documents that enable easy updating and can MoA, regions High ST accommodate different regional arrangements. • Clarify the process for updating registry books to work toward affordable mechanisms that are MoA, regions High ST accessible by local populations but also ensure a backup of critical records. Action 10: Provide clear and transparent ownership records in urban areas, thereby reducing informality. • Building on existing mapping efforts, conduct pilot exercises for urban land rights regularization in different contexts; assess the potential to integrate different land-related MoUDC High MT documents; and establish conclusive land-use records. • Assess potential benefits from unambiguous urban land records for landholders (and their MoUDC High ST willingness to pay) as well as potential for the public sector. • Disseminate good local practices in formalization of informal land holders. MoUDC High ST (on-going) Source: Authors, revised after consultation. Notes: ST: short-term; MT: medium-term. 15 1. INTRODUCTION Ethiopia’s long-term vision of becoming a middle-income country by 2025 requires sustained rapid and broad-based economic growth. The government of Ethiopia (GoE)’s new five-year plan, the Growth and Transformation Plan (GTP) (MoFED 2010), strongly reflects this point. The key pillars of its strategy toward this vision are to maintain agriculture as a major source of economic growth, create favorable conditions for the industry to play a key role in the economy, build capacity and deepen good governance; and promote women and youth empowerment and equitable benefit. Strengthening Ethiopia’s land administration system is an important element toward this vision and the listed strategic pillars. The government has recognized the importance of land administration and tenure security at least since the agriculture-led industrial development (ADLI) strategy of the government (GoE 2001)4 and has been reflected in subsequent strategic policy documents. For instance, Ethiopia’s Rural Development Policies and Strategies (RDPS), issued by the Ministry of Finance and Economic Development (GoE 2003), describes the “proper utilization, allocation and use of existing land� as 1 of 6 “basic directions of agricultural development.� Notably, the RDPS recognizes the lack of detailed legal instruments and adequate capacity for effective management of land. As a key step for consideration, the RDPS underlines the formulation of a comprehensive policy with its subsequent rules and regulations as well as the establishment and capacitating of institutions to implement the policies and directives. In 2008 the Ministry of Agriculture (MoA)5 developed the Ethiopian Sustainable Land Management Investment Framework (ESIF), which prominently features land administration in Component 2, “Improving the Recording and Demarcation of Ethiopia’s Land Resources.� The 2010 Policy and Investment Framework for Agricultural Development in Ethiopia (PIF), developed under the Comprehensive Africa Agricultural Development Programme (CAADP) framework, also stresses the importance of land administration and the need for investments in this area. The GTP itself puts land administration at the center of its chapter on “Good Governance� and calls for the establishment of a “system to register urban land� and a “system to identify and recognize ownership of land� and “proper utilization of agricultural land.�6 Such prominence of land administration in Ethiopia’s policies is no surprise, given the historical context and the multifaceted impacts that strong land administration services have on the country’s development. Just some development aspects to which land administration contributes include: 4 Therefore, the GTP is explicit on urban aspects of land administration and builds on past strategies that feature broader aspects. The five-year Plan for Accelerated and Sustained Development to End Poverty (PASDEP) (MoFED 2006) featured rural land administration. The plan stated: “As a principle, land should not be redistributed frequently. […] However, this alone is not enough. It is important, after registering the land, to provide the farmer with certified assurance to guarantee his or her user rights. Together with this, it is possible and necessary to assure the farmer that his or her land will not be redistributed for a long period (say in the coming 20–30 years).� 5 Until recently, the Ministry of Agriculture and Rural Development (MoARD). 6 With respect to land administration, the objective of the PASDEP was to provide every rural household in the 4 “main� regions of Amhara, Oromia, Southern Nations, Nationalities and People’s Region (SNNPR), and Tigray with a first-level certificate; and to issue a second-level certificate with a cadastral map to approximately 1 million households in pilot project areas. Although these targets were not achieved, significant progress was made. 16 Agricultural growth. Ethiopia’s traditional small-holder subsistence agriculture is moving toward more commercial farming and complementing small-holders with large-scale commercial farming. Improved tenure security fosters this growth and transformation. It leads to greater investments and a potentially better allocation of land and, hence, higher productivity. Experience including in Ethiopia has shown that increased tenure security leads to an expansion of the land rental market, which positively impacts agriculture (and the nonagricultural growth discussed below). Given that approximately 83 percent of Ethiopia’s population of approximately 80 million is smallholder farmers who depend completely or partly on agriculture for their livelihoods, the positive impact that tenure security has on agricultural growth is crucial. Natural resource management. Improved land administration can not only stimulate investments in agricultural land but also reduce encroachment on communal land whose boundaries and management rules are unclear. The existing ambiguities are key contributors to land degradation. The high degree of land degradation in Ethiopia threatens the livelihoods of the poorest groups of the society. Climate change is expected to exacerbate the problem. Nonagricultural growth. The World Bank investment climate assessment (2009b) identified access to land in Ethiopia as a major concern for businesses. For instance, some 56 percent of respondents of a 2006 survey indicated that their growth was significantly hampered by lack of access to land––a modest drop from 2002 but still a very significant concern. Hence, the transition of part of the population out of the agricultural sector to strengthened nonagricultural growth will depend, as one of its factors, on a land tenure system that provides security, enables efficient planning, and transfers land-use rights. Urban development. An effective land administration system is crucial for rapid urbanization. Cities depend on the system for effective urban planning, a dynamic land market, and the generation of local revenue. These elements are crucial for Ethiopia’s transformation through rapid urbanization and industrialization. Governance, including conflict resolution. Land allocation can be the object of serious disputes and corruption. Arguably, land is the number one object of conflict, including in Ethiopia. Boundary disputes and associated social conflicts can be significantly mitigated through a transparent and effective land administration system. A legal and institutional framework that secures land property rights and makes reliable information on land rights widely available to interested parties is also the basis for easy transferability and appropriate use of land as collateral, and helps avoid externalities that may arise from uncoordinated action. Rights of women and vulnerable groups. Ethiopia’s constitution and laws protect the rights of traditionally vulnerable segments of society (for example, women, pastoralists dependent on communal lands). Nevertheless, the actual implementation of these protections requires a well- organized and well-resourced land administration system. Experience, including in Ethiopia, has shown that proper land administration services can be a very effective instrument to empower women socially and economically. An effective land administration system is at the core of improving gender equality and protecting the rights of the vulnerable. The importance of an effective land administration system to various development issues is increasing. Rapid economic growth has started to transform the country and, as envisioned by the GTP, will continue to do so. Rapid growth leads to increased demand as well as social, economic, and institutional pressures on land. Functioning land planning, allocation, and markets are the bases for efficient allocation of land to smallholders and large-scale investors. These 17 processes lead to increased agricultural productivity and production. They and tenure security are crucial for long-term investments such as irrigation. Similarly, cities can no longer depend on a largely informal land market. Cities need clear and effective regulations for planning and revenue purposes as well as for rapid economic development. The government is responding to the increased demand and the need for an effective and, hence, better harmonized (national) land administration system. Particularly, the federal government is strengthening its involvement and capacity. Last year, the government, led by the then-MoARD, established a high-level task force to provide strategic guidance on land administration (MoARD 2010). In 2010, GoE also circulated a detailed investment proposal for the Ethiopia Land Administration and Land Use Development Project (LALUDEP). Its key components are designed to establish and maintain a well-functioning land administration system in the country’s rural areas.7 Significantly, the Ministry of Agriculture (MoA) also established a Directorate for Rural Land Administration and Use. The above developments make this report timely; its objectives are to provide the government with strategic options for strengthening and refining the national land administration system and for concrete interventions in this area. The report draws on inputs by the government and other development partners, and sees them as their primary target audience. It incorporates experiences and insights gained in Ethiopia and in other countries. The report highlights key issues and suggests areas in which policies, procedures, and organizational structures can be adjusted, and investments undertaken, to further develop a cost-effective, efficient, and sustainable land administration system for Ethiopia’s rural and urban areas. Land administration and related aspects of land management more broadly are enormously complex and cannot be covered comprehensively in this report. The report focuses mainly on land administration defined as “the processes of determining, recording and disseminating information about tenure, value and use of land when implementing land management policies� (UN/FIG 1999)8. In fact, the process of determining, recording, and disseminating information about tenure is at the core of this report. Given that this report has been initiated through work and discussion with the MoA, a focus is on rural land administration. Nevertheless, the report reflects the need for a national approach to land administration and, hence, close coordination between rural and urban land administration. Also, given the close linkages with land management9 more broadly, the report addresses some of these aspects as well, though by no means comprehensively. When discussing issues related to land administration in Ethiopia, we are guided by some generally principles of good land governance. They have been developed through a forthcoming World Bank (2011) study and are summarized below: 7 See Error! Reference source not found. for a summary of the preliminary design of the LALUDEP. 8 For reference: UN/FID, 1991. Report of the Workshop on Land Tenure and Cadastral Infrastructures for Sustainable Development, 18-22 October, Bathurst. www.sli.unimelb.edu.au/UNConf99/sessions/session1/bathurstdec.pdf 9 For a definition of land management see, for example, UN / ECE (1996) “Land Administration Guidelines,� which states: “Land management is the process by which the resources of land are put to good effect. It covers all activities concerned with the management of land as a resource both from an environmental and from an economic perspective. It can include farming, mineral extraction, property and estate management, and the physical planning of towns and the countryside.� 18 Legal and Institutional Framework • The law recognizes a range of rights held by individuals as well as groups (including secondary rights as well as rights held by minorities and women). • The rights recognized by law are enforced (including secondary rights as well as rights by minorities and women). • The formal definition and assignment of rights, and process of recording of rights accords with actual practice or, where it does not, provides affordable avenues for establishing such consistency in a non- discriminatory manner. • Land rights are not conditional on adherence to unrealistic standards. • Institutional mandates concerning the regulation and management of the land sector are clearly defined, duplication of responsibilities is avoided and information is shared as needed. • Policies are formulated through a legitimate decision-making process that draws on inputs from all concerned. The legal framework is non-discriminatory and institutions to enforce property rights are equally accessible to all. Land Use Planning, Management, and Taxation • Changes in land use and management regulations are made in a transparent fashion and provide significant benefits for society in general rather than just for specific groups. • Land use plans and regulations are justified, effectively implemented, do not drive large parts of the population into informality, and are able to cope with population growth. • Development permits are granted promptly and predictably. • Valuations for tax purposes are based on clear principles, applied uniformly, updated regularly, and publicly accessible. • Resources from land and property taxes are collected and the yield from land taxes exceeds the cost of collection. Management of Public Land • Public land ownership is justified, inventoried, under clear management responsibilities, and relevant information is publicly accessible. • The state expropriates land only for overall public interest and this is done efficiently. • Expropriation procedures are clear and transparent and compensation in kind or at market values is paid fairly and expeditiously. • Transfer of public land to private use follows a clear, transparent, and competitive process and payments are collected and audited. Public Provision of Land Information • The land registry provides information on different private tenure categories in a way that is geographically complete and searchable by parcel as well as by right holder and can be obtained expeditiously by all interested parties. • Registry information is updated, sufficient to make meaningful inferences on ownership • Land administration services are provided in a cost-effective manner. • Fees are determined and collected in a transparent manner. Dispute Resolution and Conflict Management • Responsibility for conflict management at different levels is clearly assigned, in line with actual practice, relevant bodies are competent in applicable legal matters, and decisions can be appealed against. • The share of land affected by pending conflicts is low and decreasing. The remainder of the report comprises four chapters: • Chapter 2 describes the historical context, reviews the legal and regulatory framework, describes the existing organizational structures dealing with land administration and 19 associated services at different levels within the federal and regional (state) governments, and reviews approaches used to define land governance and tenure rights. • Chapter 3 reviews the various aspects of the implementation of the rural land certification program, highlighting the successes as well as the subsequent difficulties, experienced in keeping land registries up to date. • Chapter 4 highlights options for key areas of modifying and improving land administration, including the legal and regulatory framework, organizational set-up and capacity-building, and actual provision of improved and cost-effective land administration services. • Finally, chapter 5 draws conclusions for the way forward. 20 2. PUTTING ETHIOPIA’S LAND GOVERNANCE IN CONTEXT 2.1 H ISTORICAL C ONTEXT Historically, land issues have been an important factor in Ethiopia’s political and economic development. The pre-1974 imperial regime supported a feudal agrarian structure, with major inequities based on ethnic identity and social class. Exploitation of the landless peasant-tenants supported the nobility, government bureaucracy, military, and church (Ayano 2009, 10). However, there were regional differences in the level of inequality. In general, the population in Amhara and Tigray with their communal or kinship land regimes had (relatively) more egalitarian access to land in, than the population in the South. Yet, even within these systems, there were significant inequities (Woldemariam 2006). As in many other countries, Ethiopia’s land-related tensions were among the key sources of social conflict and political unrest in the early 1970s, culminating in the military coup of 1974 that overthrew the imperial regime. The Derg government that emerged from the coup abolished the feudal system of land ownership. In 1975 the new government declared that all land, whether rural or urban, is state property, and organized agriculture along Marxist structures of collective farming and production quotas with a strict top-down approach to land-use planning. The government that came into power in 1991 eliminated the Marxist aspects of land use and, at the same time, maintained the notion that land is state property to which Ethiopians have an entitlement of access via usufruct rights. Article 40 (3) of the 1995 constitution states: “The right to ownership of rural and urban land, as well as of all natural resources, is exclusively vested in the State and in the peoples of Ethiopia. Land is a common property of the nations, nationalities and peoples of Ethiopia and shall not be subject to sale or to other means of exchange.� One implication of the entitlement of every Ethiopian to use land has been the periodic local redistribution of agricultural land to younger people, or to those who had a need for land. Although generally contributing to the objective of greater equity, the rural redistributions had negative impacts on land users’ perceptions of tenure security (Deininger and Jin 2006). Tenure insecurity, in turn, has been shown by numerous studies to reduce incentives for investment in land improvements, and consequently results in lower land productivity.10 Realization of these negative implications induced authorities in some areas to scale down the incidence of redistribution and to impose stricter rules defining when and where redistribution may take place. For example, in the Oromiya region, the land-use proclamation affirms no redistribution except in irrigated and so-called “unoccupied� areas. In the Amhara region, redistribution is now possible only if it is approved by more than 80 percent of the land-using households in a village. The advent of the rural land rights certification program during the past decade has been shown to have reduced farmers’ fear of future land loss through administrative intervention. Nonetheless, concerns regarding possible future redistributions within a village have not been fully eliminated (Deininger and others 2011, forthcoming). Moreover, another reason for redistribution has been the conversion of periurban land for urban residential needs 10 See Deininger and Feder (2009) for a general review of empirical evidence on this topic, and Deininger and Jin (2006) for specific evidence on the negative effects of tenure insecurity in Ethiopia. 21 and, more recently, the redevelopment of inner cities. Expropriation is practically the only mechanism for these developments. It has grown significantly with the country’s economic growth, but has likely negative impact on this growth itself as well as social development. Despite the implied constitutional censure of such exchanges, land transactions including disguised land sales have been occurring at relatively high frequency (Ayano 2009, 24). As in many countries, Ethiopia’s urban land market (whether formal or informal) has been more active than the rural market, with annual turnover estimated at 5 percent, a level compatible with major cities in some developed countries (World Bank 2007, 98). In recent years, specific laws and proclamations at both the federal and regional levels clarified that rental transactions and bequeathals of land rights are permissible (subject to certain limitations discussed in a subsequent section). The introduction of formal sanction to land transactions further enhances productivity. Rental transactions may help in transferring land to individuals or entities that have more productive uses for the land, and the ability to bequeath enhances incentives for land- improving investment.11 Land-rights concepts in areas of sedentary agriculture are reasonably similar, in a broad sense, across the different regions of Ethiopia. There is much more variation in areas in which pastoral activities predominate (Box 2.1). Box 2.1 Land Rights in Pastoral and Agropastoral Areas In pastoral and agropastoral areas, the hold of statutory law typically has been weak, because customary institutions and rules have been more prevalent. Land and water resources in pastoral areas are held communally, although in some areas, such resources are held individually, and rights are transferrable to family members. Particularly, areas of relatively higher fertility, even if only seasonally (for example, river beds), have tended to be appropriated by individuals with higher social status and utilized for seasonal agriculture. Seasonality thus could imply differing land rights regimes in the dry and wet seasons. Generally, land in pastoral areas is accessed on the basis of clan, subclan, and lineage group membership as well as social status. Under normal circumstances, access may be allowed to nonmembers, but in times of scarcity, access to nonmembers is restricted, and priority rules for access may be established. Due to the growing human and livestock populations, encroachment from sedentary populations has been increasing, as well as the award of land rights to “investors� and expropriation of land for public activities without appropriate consultation with the affected local population. Even when consultation takes place, compensation often is not provided or is inadequate (Anteneh and others 2007, 2008). Expropriation of traditional pastoral land for private investment has become a source of much concern. A comprehensive examination of land rights and resource management in pastoral areas concluded that “…there are no policies and guidelines that could be applied by any existing institution toward the management and administration of land resources in pastoral and agropastoral areas…� (Anteneh and others, 2008, xii). Under these circumstances, the access to land by groups who were of weaker social and political status to begin with (women, vulnerable people) has significantly deteriorated. Ethiopia is not unique in having a land rights system under which all land is state owned such that individuals’ and firms’ access to land is through the possession of usufructory rights only. Two notable examples of countries with similar systems are China and Vietnam (Box 2.2.2), both of which have a record of remarkable agricultural and economic performance. Obviously, the successful experience of these countries is due to many factors. However, the insight for Ethiopia is that it is possible, in principle, for the economy to perform well with a usufruct-based land rights system, provided that various relevant aspects of land policy and land 11 For example, in a sample of villages in the Amhara region in 2007, 39 % and 24 % of respondents without and with land certificates, respectively, expected some changes in the size of their holdings in coming years (Deininger 2003). 22 administration are conducive to growth. Among these aspects, perceptions of tenure security and the transferability of land rights are of prime importance. Box 2.2 Modalities and Impacts of Gradual Legal Reform in Vietnam Very similarly to Ethiopia, Vietnam historically had been characterized by highly unequal land distribution in which 2 percent of landlords occupied more than 50 percent of the country’s land, and 59 percent of the population was landless. In the early 1950s, the government redistributed land to landless tenants and rents declined. In the North, this redistribution was followed by collectivization. In contrast, in 1956 the South imposed land ceilings and rent controls and from 1970-74 a land-to-tiller program distributed land to over 1 million farmers (Pingali and Xuan 1992). After the end of the war in 1975, collectivization was encouraged throughout the country, although implemented differently across regions and generally more strictly in the North than the South. Reforms in rural areas started with kohan (contract) 100 in 1981, which stipulated in-kind labor rewards. That contract was followed in 1988 by kohan 10, which, similar to and inspired by the 1978 Chinese Household Responsibility System (HRS), provided 10–15 year land-use rights but left unresolved issues of inheritance and land transfer. To address these, the 1993 land law and its 2001 and 2003 revisions contained 2 new elements: a. They provided users with “5 rights�: transfer, exchange, lease, inheritance, and mortgage. To encourage investment, such rights were granted for 20 years on annual cropland, and for 50 years on perennial cropland. b. These laws developed regulations for land-use planning, valuation, administration, and cadastral surveying and mapping. This regulatory framework provided the basis for the issuance of more than 11 million land-use certificates (LUCs); and, even although land remains state owned, expansion of rights to transfer; and strengthened valuation capacity to provide realistic compensation. The impacts of legislated land-use rights have been tremendous. Certification fostered investment in perennial crops and, because land could be transferred without fear of loss, expansion of nonfarm activities and diversification of households’ income portfolio. Through these reforms, Vietnam transformed itself from a net importer of rice to the world’s second largest rice exporter. The country also emerged as major global exporter in high-value commodities and cash crops such as coffee, and achieved a far-reaching structural transformation. In little more than a decade (1993–2004), poverty dropped from 58 percent to less than 20 percent. 2.2 L EGAL F RAMEWORK The Ethiopian constitution stipulates several fundamental rights related to land, including the right to property (Art. 40); to development, including citizens’ right to be consulted on policies and projects affecting their communities (Art. 43); and the right to adequate compensation in case of eviction (Art 44.2). The constitution also specifies the role of the federal government in enacting laws for the utilization and conservation of land and other natural resources (Art 51.16); role of the regional governments in administering land under such federal legislation (Art 52.2(c)); and the states’ rights and responsibilities including the use land for economic objectives (Art. 89.4), the protection of the environment (Art.92), and taxation (Art.99.2 and 99.8). Ethiopia’s evolving policy, legal, and regulatory framework is specific to rural and urban land and differs across its regions. The emphasis that recent policy documents such as the ESIF or the draft GTP put on strengthening land administration in Ethiopia was made clear in chapter 1. This section focuses on the evolving legal framework for land administration. In line with Ethiopia’s federated state structure (Arts. 51 and 52 of the constitution), laws and proclamations issued by the federal government are reflected in follow-up proclamations of the regional governments, which are empowered to enact such legislation and to implement these laws. 23 The legislation in effect regarding rural land is defined by federal Proclamation 456/2005, enacted in June 2005, which replaced the earlier Proclamation No. 89/1997. Additional aspects of land policy, chiefly expropriation and compensation, are contained in Proclamation 455/2005, issued in the same month. As highlighted in the preceding section, under federal legislation, land is state property, and rural landholders have use rights that have, in principle, an indefinite duration. Nevertheless, also under federal legislation, the same land is subject to seizure by authorities for eminent domain purposes and regulated redistribution that limit the holders’ tenure security. Federal law specifically allows redistribution for land with no heirs; land considered abandoned; land that has been improved by public irrigation; and in communities that have a popular vote to redistribute land within the kebele (administrative village). Within these constraints, land-use rights can be bequeathed to family members, and any fixed improvements on the land that were produced by labor and capital are considered private property that can be sold, exchanged, bequeathed, or used as collateral. Land rights per se cannot be used as collateral, although investments into the land can, giving some flexibility particularly for large-scale investors. The federal legislation also defines the rights for eviction, expropriation, and compensation. Due to the importance of these rights––including regional legislation and implementation––this subject is dealt with separately in chapter 2.4. The federal proclamation on rural land also introduces land measurement, registration and holding certificate, land-use planning procedures and restrictions, concepts of “villagization,�12 and delivery of utilities, as well as rules for minimum holding size. The federal legislation allows a farmer to rent out the use of land to other farmers or to investors, provided that the transaction does not displace the original holder. This law prevents the landholder from leasing out her or his entire holding (although there are obvious ways to circumvent this restriction). It is required that the rental transaction be registered with, and approved by, the local authorities. Concern about land fragmentation that could lead to unviable small farms motivated federal legislators to introduce a notion of minimum size for rural landholdings. The details of minimum sizes were left to regional governments. The legislation provides equal land-use rights to women and men. Orphaned children also retain inherited land-use rights through their legal guardians. The language of the law clearly protects women’s land rights. The law therefore enhances women’s rights and standing, overturning the long tradition of discrimination against women in rural areas. That discrimination was claimed to have resulted in unfair decisions in matters of land reallocation and in the handling of women’s claims by officials and the courts in cases of divorce, widowhood, and other property settlements (Rahmato 2007, 28). The federal proclamation specifically sanctions the ability of landholders to transfer their land rights to family members through bequeathal. Family members are defined as any person who lives and shares livelihood with the landholder. It is noteworthy that the provisions in the federal proclamation regarding inheritance are not fully compatible with the civic code that governs generally issues of transfer of property through inheritance. In that code, heirs are defined through blood relationship rather than joint residence.13 This contradiction has the 12 “Villegization concentrates the residents of villagers in one central location, with the more cost-effective provision of public services one important objective. 13 Thus, the civic code would grant entitlement to land rights in land to heirs who do not live on the farm, and who are not intending to engage in agricultural activities, while the rural land proclamation apparently intended to protect 24 potential for land disputes that pit civic-code-based court decisions against rural land administration authorities’ intention to implement the land-use proclamations. The federal proclamation also sanctions the right to donate land. Given the restrictions on formal sales, the permanent transfer of land rights through donation would seem to provide an avenue for disguised sales, although the requirement to register such donations implies some measure of official scrutiny. Based on the above federal proclamations, each region (state) was mandated to formulate relevant laws and pursue the necessary actions to implement these laws at local levels. Four regions already have revised their rural land legislation and proceeded with actions to implement the revised policy (Tigray, Amhara, Oromiya, and SNNPR). Gambella issued a proclamation in 200914; Afar issued a proclamation for rural land administration in 2008; and Benishangul-Gumuz issued a proclamation in 2010. These 3 latter proclamations are not yet implemented. However, Gambella is testing the implementation of its proclamation in one kebele, and Benishangul-Gumuz has established a separate bureau in preparing to implement its proclamation. Differences in the follow-up proclamations and regulations of regions exist, as the federal legislation does not specify every aspect in great detail. In the following, a comparison is undertaken of key features of the rural land legislation in the four “main� regions that already completed the revision of their land laws after 2005.15 • Ownership. In all four regions, to conform to the constitution and the federal proclamation, land ownership is vested in the state. • Landholder rights. All four regional governments reiterate the federally proclaimed rights to use, and bequeath use rights, and acquire immovable private property (that can be sold, exchanged, and bequeathed) of the land. The Oromiya legislation specifically prohibits renting out of more than 50% of a holding and for more than 3 years. This restriction deters longer-term investment. • Redistribution. The federal proclamation specifies a number of circumstances where redistribution is sanctioned, but otherwise leaves rooms for regional interpretation. Amhara and Oromiya prohibit land redistribution other than for the exceptions specified in federal law. Nevertheless, loopholes for abuse remain. For example, renting out land by a holder who leaves for a nonfarm job may be interpreted as abandonment.16 Tigray and SNNPR do not explicitly rule out redistribution and thus imply a somewhat weaker sense of tenure security. However, a limitation on the access to land of those members of the household who lived on the farm with the deceased, and who likely are to continue farming as a source of their livelihood. 14 The proclamation is based largely on that of SNNPR, despite the agropastoral and pastoral nature of significant parts of Gambella. 15 The comparison is based on material in USAID (2008, 18–23), except when otherwise stated. The discussion of regional differences in eviction, expropriation, and compensation practices also is deferred to chapter 2.4. 16 The definition of “abandonment� may be open to different interpretations and abuse. Particularly, the taking up of nonfarm employment while maintaining low-intensity farming could be considered by officials as abandonment of the farm operation. The ambiguities entailed in these provisions introduce disincentives to undertake nonfarm employment even when opportunities exist. Only the Amhara land legislation includes a provision for redistribution of land by a majority community resolution, as mandated by the federal proclamation. 25 redistribution is implied by defining the circumstances for redistribution to land without heirs, irrigable, and vacant (abandoned) land. • Inheritance. The Tigray rural land use proclamation allows inheritance only to landless children and parents. The Amhara land proclamation is less restrictive and allows inheritance through will to other individuals. The Oromiya and SNNPR laws apparently replicate the federal law. • Land rentals. The Amhara rural land proclamation does not provide specific limits on the size of land rented out other than stating, as in the federal proclamation, that original holders should not be displaced. Rental contracts can be up to 25 years and may be renewed afterwards.17 The Amhara law requires only that rentals exceeding 3 years in duration be spelled out in a written contract and registered with local authorities. The law does not require official approval. The Oromiya and Tigray legislation allows rentals of up to 50 percent of the original landholders’ holding. Shorter contracts are renewable as well, and transfer of the contract to a third party is feasible. Legislation allows rental for up to 3 years for renters who practice traditional agriculture but significantly longer (15 years in Oromiya, 20 years in Tigray) to renters who practice mechanized agriculture (presumably in larger scale farm enterprises).18 Registration with, and approval by, local authorities is required. SNNPR allows rental transactions among peasants for durations of up to 5 years. In contrast, transactions with commercial (large-scale) investors are allowed for up to 10–25 years, respectively, for annual and perennial crops. To protect the rights of family members dependent on a farm’s income and of future heirs, all members of the original landholder’s family need to agree to the rental transaction. This feature makes it more difficult to transact in land, and therefore there is a trade-off between the objectives of equity and economic efficiency. The agreement needs to be approved and registered with local authorities. While the Amhara regulation of rentals appears to be the most liberal and conducive to the efficiency of land markets, it also entails risks. An exemption from registration of rentals of up to three years, combined with the possibility of renewing such rentals, implies that there is no actual record of the true user of the land for the purposes of land administration decisions (such as expropriation). Amhara’s legislation also engenders greater potential for disputes in the absence of written contracts. On the other hand, the restrictions imposed in the other regions inhibit the efficient operation of rental markets. The requirement of official approval of rental transactions creates opportunities for rent-seeking and may inhibit investment by commercial enterprises utilizing rural land. • Holding size. Primarily to avoid further fragmentation, almost all of the regions have determined a minimum size for the future acquisition of land. Amhara and Tigray 17 The federal proclamation does not specify limits on the duration of rental, the regional governments introduce restriction in this regard. Amhara allows for the longest duration. 18 This legislation presumably intends to attract investments into farm mechanization, which is commendable. However, it discriminates against smallholders and their investments such as in irrigation or soil and water conservation measures that might only pay off after well more than 3 years. 26 introduced a minimum size of 0.25 hectares (ha) for rain-fed land, and a maximum size for irrigated land (0.6 ha of irrigated land in Amhara, but not yet determined in Tigray). In Oromiya, the minimum sizes for rain-fed land used for annual and perennial crops are 0.5 and 0.25 ha respectively; the maximum size for irrigated land is 0.5 ha. In SNNPR the minimum size for rain-fed land is 0.5 ha, which also is the maximum size for irrigated land.19 • Gender equality. The regional proclamations reflect the principle of equal access to land-use rights for men and women and prescribe (except in Tigray) that land certificates issued by local authorities should contain the names of both husband and wife. In Oromiya and SNNPR, legislation regarding inheritance rights to land within the family has strengthened women’s status. These laws give equal rights of inheritance to sons and daughters, and widows and their dependent children can take over the land rights in case of a husband’s death (Holden and Tefera 2008). Rental agreements that transfer land-use rights require the consent of the wife, and upon divorce, husband and wife share rights over the landholding (equal rights in Tigray and Amhara). • Minors’ rights. The federal rural land proclamation recognizes the right of orphans who are still minors to receive land-use rights, to be managed on their behalf by their legal guardians. Except for Oromiya, the regional governments reflect these protections in their rural land proclamations. • Investment. The regional land proclamations are not uniformly aligned with the pro- investment orientation of the federal legislation. In Oromiya, in contradiction to the federal provision on this matter, investors may use fixed assets on the land as collateral but not as land-use rights, Furthermore, the authorities are provided the right to unilaterally revise the price of a lease on land contracted by investors from the regional authorities. Such terms likely are to reduce investors’ perceived tenure security and diminish their incentives for investment. As noted earlier, all regions except for Amhara require official approval of rental transactions, with the likely consequent delays and cost-increasing graft, thus disincentivizing investment. Urban land rights in Ethiopia are defined by two coexisting formal systems: The old “permit� or “rent� system, and the newer “urban land lease� system. Under the “permit� system, based on practices and legislation from the imperial and Derg times, landholders have the perpetual use right, which requires annual rental payments to the local authorities. Such land is not legally tradable, but the use right is automatically transferable when improvements on the land (for example, buildings) are sold (because improvement sales are legally permissible) or otherwise transferred. Indeed, “permit� land is traded frequently through this mechanism. A 1993 Urban Land Lease Proclamation (revised in proclamation 272/2002) introduced a system of land leasing that was motivated by the desire to encourage residential construction while creating a better source of revenues for cities based on property values that are closer to market valuation. The federal proclamation is reflected in parallel proclamations of regions and cities. The federal proclamation decrees that all new land allocations for any purpose, both residential housing and all pre-existing nonresidential holdings, are governed by the leasehold system. Moreover, while pre-existing residential holding rights were to be regulated by a (revised) rent 19 Motivated by concerns for equity, these rules hold for the perpetual land rights only. 27 regime, transfer of such properties through means other than inheritance or changes to nonresidential use entails the application of the leasehold regime. The proclamation allows leases of land for housing to span up to 99 years. Leases for industrial and commercial activities can span up to 80 and 70 years, respectively. Implementation of the lease system has been gradual and through various and frequently changing approaches. For example, in Amhara, only 10 cities have the leasing system introduced. In Tigray, all 12 cities with “city administration� status have the land lease, but this is a relatively recent expansion from a smaller group of cities that started leasing earlier. Leases are allocated free to government entities and religious organizations. Ways of allocating leases for other purposes have been varying over time. Thus, until recently, leases to housing cooperatives and low-income residential projects have been allocated by lottery, at very low administrative prices, and by auction or negotiation for industrial, commercial, and higher income residential purposes (Orgut 2010a, 10). However, as of March 2011, auctions and negotiated allocations have been canceled in Amhara for about two years, and leases are allocated through a lengthy administrative procedure, at administrative prices, with final approval of each transaction by the regional Cabinet. In contrast, in Mekele, allocation of land for single family housing at administrative prices has been discontinued since 2009, and land at subsidized administrative prices is allocated only for subsidized multi-apartment condominiums and in the industrial zone. For all other purposes, land leases are auctioned, and 371 plot have been leased our through auctions in the past year. While the proclamation aimed to develop optimal conditions in which leases would become the exclusive urban land-holding system in Ethiopia, the older “permit� land system remains in force, and most urban land is managed under that arrangement, which is de-facto “grandfathered.� In addition to the two formal systems outlined above, a large number of urban landholdings are illegally or informally held, having been developed and used outside of the permit or lease systems. Fransen and van Dijk (2008, 9) estimate that approximately 30 percent of households in Addis Ababa reside in illegal or informally held housing and that informality is even larger in other urban centers. All forms of land are being traded fairly actively, whether formally or informally. The illegally held land trades necessarily are handled only in an informal (primarily cash) market (World Bank 2007, 97). An intensified effort is underway in Addis Ababa to regularize informal properties. While land rights transactions among private entities (individuals or firms) generally reflect true market values, rent and lease fees to authorities generally are grossly understated. Tenure security perceptions regarding urban land differ between holders of the old “permit� lands and holders of the newer “lease� lands. Use rights on “permit� land are perceived as permanent in duration (as long as rents are being paid). Rights over leased land are perceived as being time bound. Furthermore, the courts tended to apply a civil code provision that grants full ownership of immovable assets on “permit� lands on which taxes have been duly paid for 15 or more years. This feature enhances the sense of tenure security over such land. An additional element of uncertainty applies to lease lands that are not fully paid for by the original entity or person to whom government granted the lease. The holders of such leases are considered to be in debt to the government. Consequently, what the status is of those who “buy� the lease from the original holders is not clear, as these new holders do not have a contract with government. These differences in tenure security perceptions are reflected in the fact that “permit� land carries a 66 percent–80 percent premium in the secondary land market, compared 28 to “lease� land (World Bank 2007, 100). The uncertainty also may curtail the number of secondary transactions in leases, thus causing loss of potential efficiency gains. In urban areas, the limited knowledge and understanding of the legal aspect of urban property legislation adds to the insecurity. The uncertainty is due to the fact that most urban land tenure issues, including the recognition of some of the core rights, are dealt with through less formal administrative tools such as municipal directives that rarely, if ever, are published in official instruments, or otherwise made publicly available. Tenure insecurity obviously is a significant characteristic of informally held urban properties. The absence of clear, consistent, and systematic policy and procedure regarding informal holdings was specifically identified as a key obstacle to the recognition of urban land tenure rights (Imeru 2009, 30–31). Ethiopia’s urbanization strategy and formal urban growth have been based on systemic expropriation of land from farmers in peri-urban areas. In particular, there is no mechanism of converting land rights held by farmers into urban rights. As soon as any rural territory is planned for urbanization by being included into the area covered by a Structural Development Plan (formerly, Master Plan), land holders at this territory are assumed to be subject to expropriation. This is quite unusual by international standards. Good international practices allow holders of the rural land converted into land for development to benefit from increased value of their land, but this gain is shared with the community or government in one form or another. See section 2.4 for further discussion of expropriation issues. The federal and regional governments neither always follow a proper framework nor have sufficient capacity for regulatory activities related to land administration and management. In particular, instead of following a proper cycle of (a) conceptualizing issues and possible solutions, (b) discussing them with the public and non-governmental professionals, and (c) only then codifying these solutions in laws and regulations, the steps (a) and (b) are often insufficient. This leads to flawed regulations, both conceptually and technically. Furthermore, the mindset typical for centrally planned economies still often prevails in the land administration and management arena. In particular, the cost and practical feasibility of implementation seems to be not always considered by lawmakers. The case of the very complicated land leasing system discussed further is an important illustration of this weakness of the current legislative practices. 2.3 O RGANIZATIONAL S ET- UP The responsibility for land administration in Ethiopia is divided among different institutions responsible for rural and urban land. In addition, the Ethiopian Mapping Agency (EMA) plays an important role in land administration (Box 2.3). Responsibilities also are shared between the federal and regional level, with large differences in structure and capacity between the regions. 29 Box 2.3 Ethiopian Mapping Agency The EMA (www.ema.gov.et) is an autonomous federal agency under the Ministry of Finance and Economic Development (MoFED). EMA provides needed inputs for the land registration and certification process and for land-use planning. These inputs include the (a) the establishment of geodetic control networks, consisting of points with recorded coordinates throughout the national land space. These points enable the referencing of field measurements to the network so that information from multiple local surveys can be accurately located; (b) preparation of base maps for rural and urban areas (with scales appropriate for typical uses of such maps in the different areas); (c) establishment of standards for surveying and mapping, and monitoring their application; and (d) certification of professionals engaged in mapping and surveying, and capacity building to promote such skills. EMA operates throughout the country but does not have regional offices. Its capacity to carry out its various functions is severely limited due to lack of resources (Orgut 2010a, 9). At the federal level, rural land matters are handled by the MoA. In the past, the MoA did not have a specialized unit dealing with land administration matters. However, it recently established a Directorate for Land Administration and Use, which reports to the State Minister for Natural Resources. The key responsibilities of this directorate are to implement the Rural Land Administration and Use Proclamation by providing professional support and coordinating competent authorities. The directorate links the work at the federal level with that at the regional level and provides inputs for policymaking to advance the harmonization of land administration. In addition, the MoA has an Agricultural Investment Land Administration Team under its Agricultural Investment Directorate (Box 2.4). The existence of two agencies with similar subject matter raises concerns about coordination. MoA’s professional capacity––that is, its skilled and specialized experts––is both expected and required to be strengthened to lead a national program in land administration. Box 2.4 Administration of Land for Large-Scale Agricultural Investors The administration of agricultural land for large-scale investors is carried out by separate departments or agencies than those responsible for general rural land administration and use, both at the federal level and in many regions. For instance, MoA’s Agricultural Investment Directorate is responsible not only for attracting investors and providing investment support, but also––through its Agricultural Investment Land Administration Team––for the overall administration of all agricultural land plots exceeding 5,000 ha, including clarifying boundaries, holding records, and valuation. In several regions, the separate entities for the administration of agricultural land for large-scale investors and that of general rural land do not even reside within the same agency. General rural land administration might be the responsibility of dedicated land administration bureaus while agricultural land for investors (below 5,000 ha) is the mandate of the regional Investment Office or the Bureau of Agriculture. These organizational arrangements imply significant coordination challenges and fragmentation of land administration capacity. Regional land administration systems vary across Ethiopia. For instance, rural land administration matters in Amhara are handled by the Bureau of Environmental Protection, Land Administration and Use (BoEPLAU), reporting directly to the regional government. Tigray also has an Environmental Protection, Land Administration and Use Agency (EPLAUA). It is accountable to the regional Bureau of Agricultural (BoA), but the agency has an independent budget. In 2009 Oromiya established (by Proclamation 147/2009) a Bureau of Land and Environmental Protection (BoLEP), reporting to the head of the regional government. The 30 BoLEP aims to create a single executive organ both to administer and regulate the urban and rural land, and to prepare land-use plans for the region. In SNNPR, the responsibility for rural land administration has seen frequent changes. It evolved from a Land Administration and Use Team under the head of the BoARD to a separate environment and land administration bureau, and, very recently, to a newly established Environment and Natural Resources Bureau. The details and specific status of the land administration under this latest set-up are yet to be worked out. The administrative structure and capacity of emerging regions are being built but are still in their infancy. Benishangul-Gumuz has had an EPLAUA for long time, but the agency did little work on (rural) land administration. However, the recently enacted regional land proclamation upgraded the authority to a Bureau of Environment, Land Administration and Use. The bureau has a more concrete mandate, including the support of administrative structures at the woreda and kebele levels. Gambella very recently established a Bureau of Land Use, Land Administration and Environmental Protection, although detailed structures are still unknown. The Afar and Somali regional states are making efforts to develop institutions. At present, these regions’ efforts are being coordinated by the Bureau of Pastoralist, Agriculture and Rural Development and Bureau of Livestock, Crop and Rural Development, respectively. Urban land is handled at the federal level by the Ministry of Urban Development and Construction (MoUDC),20 created in 2005 to implement the government’s urban development policy. In its role as the coordinator of the urban affairs of the country through its Land Development and Administration Department, MoUDC supports urban local governments with policy, planning, capacity building, and guidelines. The department’s resources are limited: only 5 professionals work on a combination of land information, land development, and land administration. Consequently, support to local city governments is limited and primarily on request. These local governments also receive assistance from regional governments’ relevant offices. The MoUDC’s main focus is to support physical and structural urban planning in fast- growing cities. It accords lower priority to land administration matters (Orgut 2010a, 9). A very recent business process re-engineering (BPR) study of the ministry identified weaknesses in putting details of land development and administration within the mandate of the ministry in a relevant and effective institutional arrangement. Such an arrangement would help to effect the enacted lease, compensation, and other legislation; support the regions with capacity; and support the focus on and understanding of the significant relevance of land to other duty processes (land-use plan, housing development, policy and law, and good governance). The actual implementation of the various functions that affect urban land and real estate markets is handled by city and municipality governments. This urban administration is carried out in accordance with relevant regional proclamations, with the exception of the cities of Addis Ababa and Dire Dawa, which have regional status.21 City governments have the authority and mandate to deliver numerous services, including land administration services, and to raise local revenue. The institutional arrangements have different shapes, often are complex, 20 Until recently, the Ministry of Works and Urban Development (MoWUD). 21 These are urban settlements recognized as “cities� by the respective regional governments. Ethiopia has close to 100 such cities. Smaller urban settlements that are not designated as cities (generally with populations of up to 22,000) are not autonomous; their administrations report to woreda councils. In such urban settlements, land administration matters are handled by the respective woreda offices. 31 and are changing in many cases. For instance, the Addis Ababa City Administration has three relevant institutions: those responsible for land development, for planning and information, and for building permits and land administration (structured up to the kebele level with specific duties); it is now in the process of establishing a separate registration agency. Functions performed by (non-regional) city governments and the instruments available to them are not aligned yet with the tried and tested international practices. In particular, in the absence of an independent system of registering or recording real estate transactions, city administrations are in charge for recording transactions, certifying property rights, and maintaining records and files. Moreover, city administrations are directly involved in routine real estate transactions: each transaction might require “no objections� from several departments. In some municipal governments, the same agency––the urban planning and land administration department––is responsible for both land management (making decisions on land use and controlling development) and land administration (keeping records, underpinning secure tenure, and facilitating transactions). Consequently, the agency often is ineffective at doing both and, perhaps more importantly, provides poor service to the property-owning citizens of the city (DHV Consultants 2006b, 3). On the other hand, what city governments are permitted to decide regarding land management and land administration is stiffly limited by the federal and regional levels of government, comparing to their counterparts in other countries. In particular, city governments, even those of such advanced cities as Mekele, do not have direct control over numerous parameters that are local by their nature, such as starting prices on auctions, rates for annual “permit� payments, parameter of land lease contracts, etc. As Ethiopia’s land administration system is still evolving, it is of interest to observe how land administration is handled in several other transitional and developing countries (attachment 1). In particular, it is noted that typically a unified land agency handles land administration for both urban and rural land. Also, the common international practice is that local governments are not involved in concluding and authenticating real estate transactions.22 In some countries, for historical reasons, cadastral functions and title registrations are handled by separate agencies, although this is not considered best practice (Williamson 2000). 2.4 E XPROPRIATION AND C OMPENSATION Current legislation and regulation for compensation has shortcoming. Legislation has intended to introduce some fairness and uniformity in compensation for land rights being revoked to facilitate government-authorized take-over of land. Nonetheless, a 2007 comprehensive assessment of compensation practices concluded that: “…the problems associated with the legal, technical, institutional, and financial aspects of rural land valuation and compensation practices are numerous, and the practices are full of inconsistencies, unfairness and are not standardized. The valuation methods and compensation procedures vary intra- and inter-regionally, for that matter even within a given woreda due to various causes and reasons. Subjectivity and inconsistencies in valuation and compensation are apparent even for land appropriated for the same 22 Except a few countries in the Middle East. 32 purpose, among and at times depending on the institutions involved and the committees established.� ––Anteneh and others 2007, xvii Ethiopia’s federal legislation dealing with expropriation for eminent domain purposes (Proclamation 455/2005) has expanded the previously existing definition of eminent domain by authorizing expropriation for “public purpose.� This definition is intended to better promote development.23 The interpretation of what is a public purpose that is better for development (and that therefore is eligible for expropriation) is relegated to local officials (Ayano 2009, 33).24 Urban expansion in Ethiopia entailed systematic expropriation of periurban land and reallocation of it as housing plots to urban dwellers from special waitlists. In recent years, there were several cases in which local authorities expropriated periurban land for residential development by private entrepreneurs. In the authorities’ perception, the land expropriated from farmers serves the urban community’s need for housing (World Bank 2007). The federal expropriation and compensation law underlines the urban structural plan or development plan as its basis for decision to expropriate, thereby leading to a dominance of urban interests, specifically the new land users. In general, the definition of eminent domain under proclamation 455/2005 “is quite broad, so much so that if the authorities so desire, they may be able to deem any activity as serving the public purpose,� thus facilitating extensive expropriation (Anteneh and others 2007, 20). An alternative and stricter approach to expropriation, which seems to strongly promote development, is applied in Peru (Box 2.5). The regional follow-up legislation differs across the regions. The regional land administration proclamations are diverse in reflecting the issue of public use/purpose. To better define the term and focus it on potential social and economic services that may be directly or indirectly benefiting the locality, the legislation of Amhara and SNNPR prohibits the expropriation of land without the landholder’s consent, except for cases of public interest. In Oromiya, a landholder’s use rights are superseded by more important public uses for the land, but eviction for the purpose of settling a liability or an execution of judgment is prohibited. The Tigray law does not state a clear prohibition on expropriation; neither does it define “public use� other than to specify the case of periurban land needed for city expansion. Federal legislation requires compensation but vaguely and without a requirement to rehabilitation measures. By law, landholders whose land is expropriated by authorities to serve public use needs or for private development perceived to be in the public interest (for example, residential complexes) are entitled to substitute land or compensation.25 The existing federal legislation26 decrees that the original holders of expropriated land be compensated at a rate determined by the federal government if the land is taken for a federal purpose; or at a rate determined by regional authorities if the land is taken by regional or local authorities. Alternatively, expropriated landholders may be given other land in substitution. Federal laws 23 The earlier proclamation that dealt with expropriation (401/2004) confined the taking over of privately used land by government to “appropriation of land for government works,� where government works were defined specifically by examples as highways, power plants, airports, railways and other similar public infrastructure activities. 24 The potential for abuse is not unique to Ethiopia. In 2005 the USA Supreme Court has ruled for a more flexible definition of the eminent domain concept, allowing expropriation by local authorities to facilitate large scale private enterprise. See Ryskamp (2006). 25 As spelled out in Proclamation 455/2005 and further detailed in Council of Ministers Regulation 135/2007. 26 The Federal Proclamations Nos. 455/2005 and 456/2005 are not consistent between themselves on the matter of compensation (USAID 2008, 20). 33 further require that the compensation rate take into account the existing improvement on the land taken. However, the regulations and practice related to expropriation do not stipulate any rehabilitation measures (i.e. relocation assistance, restoration of the livelihood) that are recommended by all international good practices. Moreover, expropriation to enable large-scale private development authorized by authorities does not entail a negotiation between the private developers and the smallholders affected, of which the final rate is more likely to reflect the landholders’ forgone future income stream. Rather, the compensation is handled by authorities, disadvantaging the smallholders (Rahmato 2007, 16). An assessment of the levels and practices of compensation actually received upon expropriation over the past few years indicates that the system generally is characterized by unfair and inconsistent valuation methods, leading to inadequate compensation (Anteneh and others 2007, xvii). Box 2.5 Strict Limits on Expropriation and Devolution of State Land in Peru To uphold private land rights and prevent government from abusing its power, Peru’s constitutional rules tightly circumscribe cases in which expropriation can be used. Art. 70 of the constitution stipulates that expropriations can be carried out only for reasons of national security or “public need� (for example, to build a road or bridge with no clear beneficiary). The expropriations law clearly states that expropriations are void unless the state is the direct beneficiary. Public scrutiny and debate of individual expropriations are ensured by the requirement that any case of expropriation must be authorized by Congress in a law explicitly spelling out the future use of expropriated goods. To ensure an impartial and realistic valuation, property values are determined in a court proceeding. Expropriated owners have the right to cash payment of the land’s market value plus remedies for any damages. To avoid abuse, mandatory deadlines have been established as well. Expropriation orders will lapse automatically after 6 months if the judiciary process to assess values has not started by then; and after 24 months if court proceedings are not concluded by then. Moreover, if within one year of the conclusion of the court process the expropriated property is not used for its planned purpose, it automatically will revert to the original owner. These safeguards against inappropriate expropriation has not undermined investment. To the contrary, it attracted investors who transformed the country into a key agro-exporter of high-value crops, creating significant employment. The investment promotion agency (Proinversión) defined mechanisms to attract investors depending on whether the project is initiated by the government or by an investor. If a (central ministry, regional, or local) government agency determines that carrying out a project on state land would be desirable, it will ask Proinversión to initiate a process to regularize land rights and to determine the nature of pre-existing claims that may need to be respected or cleared, and the type of land rights that can be granted to private investors. Any illegitimate claims are cleared through the proper channels. If there are legitimate claims, the rules for expropriation and compensation are followed. Any technical defects of the title also will be cleared in the process, for example, if a tract of public land is dispersed in many records with poor descriptions, a correction in the property registry will be made. The intention of divesting the land then is published in the official gazette, local and international newspapers, and a government website. The terms of bidding (that is, minimum investment required and minimum bid price for the land) are published for a minimum of 90 days, or longer if the project’s complexity warrants it. Before bids are accepted, bidders must prequalify by proving that they have liquid assets to cover at least 60 percent of the minimum bid price plus the desired amount of investment on the land. For prequalified bidders, bids are assessed and ranked by the offered price plus the amount of projected investment. Monetary offers then are presented, and a winner is declared. Before the signature of the contract, the land must be paid for and a letter of credit covering the amount of the proposed investment deposited with the government. If a piece of land is identified by a private investor, a similar process applies. In this case, the investor must present a detailed business plan including the amount of proposed investment and price for the land to a board of technical specialists from the public and private sectors. If the proposed project is deemed viable and does not conflict with existing regulations, the proposal is published for a minimum of 90 days to enable other potential investors to offer to carry out the project. If any investor comes forward, a public bidding process as described above will be initiated. If, during the 90-day publication period, no one has shown interest in the project, the investor is allowed to proceed as originally proposed. 34 Compensation also differs significantly across regions. The Amhara proclamation provides a formula for the rate of compensation that is the same regardless of whether the payer is government, private organization, local community, or any other entity. In Oromiya, holders of expropriated land are to receive compensation that recognizes both existing improvements and the value of future benefits lost through the award of equivalent land to the extent possible, and monetary compensation otherwise. The treatment of holders of expropriated land evidently is more generous in Oromiya, compared to the other regions, in which compensation for costs of displacement and rehabilitation is not considered. The SNNPR makes land substitution mandatory where feasible and follows the federal compensation rate. The inconsistencies among regions and between the regions and the federal government are reflected in different rates of compensation for similar land. Proclamation 455/2005 does not specify which authority has the final say in determining the compensation levels to be paid to those affected by land expropriation. The observed practice seems to be that initial compensation assessments for land taken to accommodate woreda and kebele infrastructure and development projects are handled by committees set up by local woreda and urban administrations, as required by article 10 of the federal proclamation 455/2005. (In some occasions, zonal administration may be involved.) The woreda committees dealing with initial evaluations are established ad hoc, and there is no guidance on how to compose their membership. Consequently, in different locations and over time, there may be different skills mixes and administrative representation on these committees. Particularly, there is no requirement for them (nor is it the common practice) to include within their ranks civil society representatives, community leaders, or anyone who could represent the interests of the affected landholders. Often, these committees have no comprehensive directives or guidance on what kind of formulae to use or, more generally, on how to conduct the valuation. Except for the SNNPR, once initial valuations are completed, the ultimate decision on compensation levels to landholders affected by woreda or kebele projects is retained by regional government offices (Land Administration and Utilization Authority in Amhara and Tigray, and Investment Commission for private investment projects only in Oromiya). On expropriation, landholdings already incorporated within urban limits (even if they are being used for agriculture) are valuated by urban authorities, who determine the level of compensation. The broad applicability of the expropriation law and the unfair compensation imply reduced tenure security, with obvious negative implications not only for equity but also for investment incentives to smallholders. There also are some (unconfirmed) claims from municipality officials that, when expropriated-land users can abuse the system, they also may seek excessive compensation, hence hampering development. Expropriation for urban growth, as it has been practiced, is not sustainable economically and socially and needs to be replaced by new approaches. As illustrated by examples in section on Land Management in chapter 2.6 below, the practice of land expropriation for urban growth is not sustainable: on one hand, city administrations cannot afford paying compensations even in the amounts deemed unfair by replaced rural tenants; on the other hand, mass replacement of sitting land tenants has high social cost. Hence, there is an acute need for establishing new policies and practices that would reduce expropriation.27 The problem should be tackled from three main directions. First, a clear legal mechanism should be established that 27 It is noteworthy that in an opinion survey of urban government officials in Amhara and Tigray in 2006, about 1/3 (14 out of 42 people) said that land expropriation in their city or around could be reduced or avoided. 35 would automatically convert legitimate land holding rights of rural tenants into urban land rights when a rural territory is planned for an urban expansion. Second, new policy guidance should specifically call for planning urban expansion that minimizes expropriation. Second, land development techniques based on voluntary participation of sitting tenants (such as the land re- adjustment technique or “land pooling�) should be deployed, instead of massive expropriation. Monitoring of the magnitude of land expropriation and assessment of its implications for people affected would be an important first step to assess and improve the situation. While it is broadly recognized by officials at regional and local governments that expropriation practices have been unfair and often devastating for displaced farmers, no authority collects unified statistics on how many households are displaced annually or losing their livelihoods in each region as a result of the current urbanization practices. There are no studies of how displaced people have been coupling either. Overcoming these shortcomings would be the basis for improving the regulations and practices. 2.5 D ISPUTE RESOLUTION AND CONFLICT MANAGEMENT Article 12 of the Federal Proclamation on Rural Land Administration and Use (456/2005) provides that in cases of disputes over rural land-use rights, efforts should be made to resolve the dispute through discussion and agreement. If these avenues do not resolve the dispute, an arbitration body is to be set up. Its composition is to be decided by the parties to the dispute or in accordance with rural land administration practices as established by regional governments. In following up on the federal proclamation, the regional proclamations pertaining to rural land specify mechanisms for dispute resolution that entail arbitration as part of the available procedures. There are some differences among the regions. The Tigray Proclamation No. 136 (Article 28) allows the kebele Land Administration Committee (LAC) to handle and decide land dispute cases. The legislation in Amhara, Oromiya and SNNPR does not empower their kebele LACs to undertake decisions. Rather, the land dispute cases are brought before the LACs to facilitate and moderate the negotiation and arbitration through a local elder selected by the parties. It was observed that, in Amhara, the practice is that land disputes first are brought and resolved by the LAC, which always has traditional leaders in its ranks, thus de facto following the traditional mechanisms of dispute resolution through community elders (Rahmato 2007, 10). The Oromiya proclamation (130/2007, Article 16.1) specifically provides that, to recognize it formally, the outcome of the arbitration should be registered at the kebele land administration office. The other regions do not mandate such formalization, which likely leads to lack of uniformity in the way that different localities handle dispute outcomes. Amhara has empowered kebele-based Social Courts to deal with, among other things, land disputes whose monetary implications do not exceed ETB 1,000. The judges in these courts are nominated by the kebele executive and approved by the Council. When the monetary magnitude at stake exceeds Social Court authority, disputants may take cases directly to a woreda court. A survey in Amhara, which likely represents a more general attitude among rural residents, reveals that 80 percent of the respondents prefer to take land disputes to arbitration by customary procedures (elders, religious leaders). As a second option, 94 percent stated that they would take their cases to the kebele office (where presumably the case is referred to the LACs that serve as a permanent kebele land administration organ). As a 36 third option, 67 percent referred to the Social Court (Rahmato 2007, 22). Therefore, the citizens evidence a clear preference to avoid higher levels of formality, for which the monetary and time costs are high and act as deterrents; and to utilize customary low-cost approaches. Generally, the reliance on kebele-level dispute resolution procedures seems to have made the process simpler, fairer, and faster.28 The preference for traditional local mechanisms seems particularly strong in pastoral areas (Box 2.6) Box 2.6 Dispute Resolution Mechanism in Pastoral and Agropastoral Areas The traditional mechanisms for dispute resolution and conflict management in pastoral and agropastoral areas gradually have lost their effectiveness––at the time of intensification and increased frequency of conflicts over resource use and control due to the pressure of growing populations and ever-increasing scarcity of resources. Causes may be both the fraying of societal cohesion and undue intervention by local government administrations that further undermines the authority of customary leaders. Many of the disputes are occurring between community members and outsiders due to encroachment on common property land that traditionally “belonged� to the community. The customary leaders’ informal authority does not hold as much weight with outsiders, and their power is further eroded by interference from formal (government) authorities. As a consequence, in pastoral areas, traditional conflict resolution mechanisms have become inadequate (Anteneh and others 2008). While the conflicts referred to above are among individuals, another source of disputes are cases in which government (represented by officials at various levels) makes land- related decisions that negatively affect existing landholders’ rights (for example, expropriation, land exchange, and compensation levels). Federal and regional governments have not established a transparent dispute resolution mechanism or an appeals procedure for these cases. It is only with respect to compensation levels (but not with respect to the expropriation itself) that a procedure to deal with grievances has been formulated. This procedure (Art. 11, Proclamation 455/2005) indicates that, in the case of the expropriation of rural land, dissatisfaction with the amount of compensation can be referred to a regular court of competent jurisdiction. The decision of such a court can be appealed further (within 30 days from the transmission of the lower court decision) to a regular appellate court. In the case of expropriated urban land, the prescribed appeals procedure relies on specialized administrative committees (although the courts could be used if such a committee has not yet been established). Utilization of the appeals procedures is subject to tight deadlines and presumes that the aggrieved landholders are familiar with the procedure and its deadlines, which familiarity most likely is not the case. Except for Amhara, regional proclamations have not explicitly specified regulations that follow the procedures outlined in the federal legislation. The Amhara regulation pertaining to rural land deals with the grievance process in a manner that deviates from the federal procedure. The Amhara regulation provides that disputes regarding the level or manner of compensation for expropriation may be submitted to a Compensation Grievances Investigation Committee within 15 days from the expropriation notice (the federal proclamation did not mention such a committee for rural lands). The committee is appointed by the same government that initiates the expropriation act so is not fully independent. Decisions of this committee (or decisions of the woreda, in case there were no such committee) can be appealed to the woreda court rather than to the court with appropriate jurisdiction. In the absence of a free legal advice facility sponsored either by government or a nongovernment organization, the actual utilization of such 28 As revealed in a survey of 160 respondents commissioned by USAID 2008, 33. 37 dispute resolution procedures by poorer rural smallholders is doubtful.29 Smallholders tend to be less informed about legal provisions and protections, and suspicious of the formal system. The appeal procedure for the level of compensation for expropriated land is different in the case of urban areas. First-instance complaints and grievances in urban areas are required to be submitted to administrative committees that will be established for such purpose. Only if these committees have not yet been set up, the grievance can be submitted to a regular court that has jurisdiction, with the right to appeal further, as in rural areas. As in the Amhara regulation, the administrative committees are appointed by the same government entity that initiates the expropriation, which hampers impartiality. A detailed review of actual compensation practices that are applied across Ethiopia concluded that even although the federally prescribed principles and procedures for compensation and associated dispute resolution are not explicitly discriminating against the poor, the deficient implementation and the actual practices most often result in anti-poor outcomes (Anteneh and others 2007, 35). This unfair treatment is aggravated by smaller and poorer landholders’ lack of familiarity with legal protections and appeals options. As pointed out earlier, there is no appeal option to the expropriation decision itself. The eminent domain legislation leaves it completely up to officials to interpret which would be a better use of a particular parcel to justify expropriation. This legislation and the generally inadequate levels of compensation are the sources of much discontent (Anteneh and others 2008, 24). 2.6 L AND -U SE P LANNING Legal and strategy foundations exist to promote the preparation of land-use plans. Articles 13 and 16 (3) of the federal Rural Land Administration and Use Proclamation No. 456/2005 call for preparation and implementation of land-use plans and establishment of land administration and use information systems. The Afar, Amhara, Benishangul-Gumuz, Gambella, Oromiya, SNNPR, and Tigray regional rural land administration and use laws have embedded a number of provisions that call for development and implementation of land-use plans. Issued by MoARD in November 2008, the Ethiopia Strategic Investment Framework (ESIF) for sustainable land management (SLM) further details many aspects of rural land-use planning at the national, regional, and local levels. ESIF suggests a participatory approach that involves communities and local stakeholders more extensively than in the past. However, comprehensive land-use plans for rural areas are not yet available in Ethiopia, and socioeconomic and political factors often have not been properly taken in account. Local authorities’ or even citizens’ participation in land-use planning has been rare.30 At times, this approach led to the proliferation of unfeasible plans that could cause negative social consequences. A balance needs to be struck between the considerations that are derived from physical and environmental factors and the socioeconomic contexts, and the interest of local communities and individuals. Based on a recent review, Imeru (2009) pointed out that, in the absence of land-use plans, land use is changing in contravention of existing regulations; and that 29 In other countries that have introduced significant changes in land administration, it has become good practice to provide free legal advisory services to protect the interest of the poor (World Bank 2009). 30 During the Derg regime, land-use planning originated from a top-down planning approach. Translating soil surveys and land capability assessments, it tended to focus on the relationship between land use and the environmental parameters of the geographical unit. 38 the share of land set aside for a specific use that is used for a nonspecified purpose is greater than 50 percent. Similar to the case of rural land-use planning, the traditional approach to urban planning in Ethiopia was centralized and inflexible, even although it generated professionally good- quality city plans. Plans were produced for all cities by the National Urban Planning Institute, and they could not be amended. While these plans were highly prescriptive, detailed, their preparation often did not entail much participation by city authorities or feedback from community representatives and business interests. Implementation of such plans sometimes was followed rigidly, to the detriment of communities’ interests. More often, due to lack of local “ownership� of the plans and their conflict with effective demand for land, real land uses deviated from some aspects of the plans, creating inconsistencies, inefficiencies, and placing land users into “gray areas� of not completely legal activities. In general, the prescriptive and over-detailed plans, while not responding to effective economic demand for various land uses, have been reducing the economic productivity of urban land and thus reducing the contribution of urban land in poverty reduction. Moreover, excessive prescriptions and limitations with respect to what can be done on the land do not represent any real public interests. A typical example is the spatial division of industrial zones into detailed subcategories (e.g., manufacturing & processing; handicrafts, garages, & workshops; silos & warehouses), which induce considerable economic and administrative costs. Moreover, such excessive description leads to public losses by reducing land market value and, hence, potential revenues from land auctions. More recent city proclamations by regions, and federal charters issued by federal authorities, led to a greater decentralization in the urban planning process. The federal Urban Planning Proclamation No. 574/2008 provides that three hierarchies of plans shall be considered, namely, the National Urban Development Scheme, Regional Urban Development Plan, and Urban Plans. The proclamation also recognizes two types of urban plans: the City- Wide Structural Plan and Local Development Plan (LDP). The federal urban land lease proclamation states that urban land shall be permitted to be held by lease in conformity with the urban plan, where such a plan exists (Proclamation No. 272/2002, Article 4/1/a). Thus, the legal framework in Ethiopia envisages urban development to be guided or controlled by urban plans. It is not clear at this moment whether the new urban planning system established for urban areas by the federal Urban Planning Proclamation No. 574/2008 and respective implementation regulations and detailed manuals (see the 2d section below) will be leading to land use planning practices that would stimulate economic development. In particular, for sustainable development of cities and in order to stimulate legal / formal private investment and construction, strategic changes in urban planning are needed and they should aim at: • Increasing flexibility of land uses by permitting multiple compatible land uses (including mixed-use) and relaxing other restrictions in structural development or master plans and all other spatial planning documents; mixed-use should be permitted and encouraged as a broad planning principle, especially in central parts of cities; • Planning territorial development according to demand signals. This can be achieved by establishing simple monitoring of territorial aspects of private construction activity, in order to plan and support urban development in these directions. This kind of monitoring requires, first of all, mapping (approximately) locations of properties for which building permits were issued. 39 • Planning territorial development in a way that reduces land expropriation (including from farmers and informal settlers) to an absolutely unavoidable minimum as discussed in section 2.4. Promising improvements of urban planning and land management are emerging. For instance, based on its flexible 2008 Structural Development Plan Mekele has legalized 12,800 informal land tenants through land re-parcelization and community participation. It is also implementing a Local Development Plan in the central city that balances the interests of current holders of small land parcels with the need of land assemblage and re-parcelization for renewal of the city fabric. Regional proclamations empower urban administrations to undertake their own planning to the extent their capacities allow such activities. Otherwise, cities were allowed to contract out the preparation of plans. In either case, the process of planning is now their responsibility, thus ensuring adequate local input and more realistic plans. In recognition that fast-growing towns require a more flexible approach to planning their spatial layouts, along with the local autonomy in planning, strategic planning has been introduced at federal, regional, and local levels. A number of important changes provided the set-up for the new approaches. At the federal level, the renamed Federal Urban Planning Institute (FUPI) shifted its activities from generating urban plans to formulating policy and developing private sector planning capacity. Similarly, the federal government has shifted the focus from preparing master plans to providing national- level institutional frameworks. GoE policy aspires to utilize regional development plans to strengthen urban-rural linkages and thus improve economic performance. The federal level has issued detailed manuals to support the implementation of the new legislation by urban centers. Particularly, the manuals provide guidance for the preparation of city-wide Structure Plans and Local (neighborhood) Development Plans. The latter provide a mechanism for bringing the planning process to the smallest spatial units within the city. The former facilitate, among others, rural-urban coordination that takes advantage of socioeconomic and physical synergies and aims to minimize conflicts due to urban encroachment in rural hinterlands. It would have been useful to augment these mechanisms with an institutional arrangement for appealing plans, such as federal, regional, and city-level appeals committees, so that conflicts and inequities can be reconciled (World Bank 2007, 165). While the reformed institutional framework now exists, and general guidance is available, a major capacity building effort is needed to transfer the techniques for implementing city- wide structure and local planning activities, to transform the attitudes of urban planners from prescriptive and normative to one attuned to the needs of the urban economy, and to instill the confidence required for officials to be able to take initiative and adapt manuals and instructions to their cities’ specific circumstances. In particular, urban planners at all three levels (federal, regional, city) need training on basics of land economics and implications for the planning practice. The largest city in the country, Addis Ababa, has a master plan that was prepared in 2000 by revising the 1986 Master Plan and enacted into law in 2002. The City Government also completed the study of over 40 Local Development Plans, and three plans are being developed. Although there are efforts to enforce these land-use plans, they are not effectively controlling urban development in Addis Ababa. According to a recent review (Imeru 2009), the development of the central and older parts of the city occurs ad hoc with infrastructure struggling to catch up. In periurban areas of the city, development almost invariably occurs well in advance of infrastructure. The same review observes that effective control of urban 40 development by land-use plans in Dire Dawa, Adama, Gondar, and Mekele is similar or even worse than in Addis Ababa. Although these cities had initiated master plan preparation, none of the cities yet has been effective in controlling urban development through such plans. Another key constraint is presented by the inadequate land information system. There is a lack of land inventories, land maps, and uniform plot identification systems. In their absence, the formulation of local plans is difficult. New urban planning policies should encourage planning spatial growth and redevelopment that reduces the need for expropriation. First, mobilization of inside-town vacant or underutilized land before going for peri-urban expropriation can provide some temporary relief. Second, urban planners in charge of structural development plans (master plans) and other spatial planning documents must be required to plan urban expansion with “expropriation impact� effects factored in (which has not been practiced). Finally, city governments should be prohibited to displace existing legal land users on the ground of non- conformity with new urban plans.31 2.7 O THER L AND M ANAGEMENT A SPECTS AND L AND -R ELATED R EVENUES Ethiopia’s rural land resources, critical to the economic and social development of the country, are seriously degraded.32 The need is urgent to reverse the serious levels of land degradation through promoting and scaling up successful SLM technologies and approaches. Specifically, to accomplish this objective will require (a) acquiring and using the required knowledge and technological advances; (b) addressing policy and legal gaps; and (c) strengthening relevant institutions. The improvement of land tenure security through further strengthening land policies and administration is an important incentive for boosting farmers’ investments in SLM practices. Additional improvements in the arrangements for land administration, particularly the provisions within the existing legislation to establish better land registration and certification, are important elements to implement better tenure security, and thereby to facilitate wider adoption of SLM practices among land users. To guide and implement SLM activities in Ethiopia, MoARD developed the ESIF noted earlier. The ESIF explicitly recognizes that secure property rights in land are essential for a successful strategy implementation (MoA 2008, 36–37). The framework distinguishes between the densely populated highlands, in which individual rights are more relevant; and the less populated areas, in which communal rights traditionally have prevailed. To implement the required policies and investments that address tenure security, the second component of ESIF is focused on improving land administration and tenure, which entails seven sub-components. They are (a) completing and updating the first-stage land certification process; (b) registering traditional use and property rights and responsibilities for the use and management of communal 31 Moreover, street addressing system still does not exist in many places, which reduces efficiency of many municipal services. Simple and inexpensive systems of introducing street addressing exist and training can be obtained through the World Bank Institute’s e-learning system. See Farvacque-Vitkovic, Catherine, Luien Goden, Huges Leroux, Florence Verdet and Robert Chavez, 2005. Street Addressing and the Management of Cities. - World Bank, Washington, DC. 32 Ethiopia is considered to be one of Sub-Saharan Africa’s countries most seriously affected by land degradation. By the mid 1980s, some 27 million ha, or almost 50 percent of the highland area, was deemed to be significantly eroded. Of this land, approximately 14 million ha are seriously eroded and over 2 million ha beyond reclamation (FAO 1996), A more recent study (World Bank 2006) estimates the total costs of land degradation in Ethiopia to be between 2 percent–3 percent of GDP. 41 land resources; (c) developing and initiating second-stage land certification; (d) building up the capacity for land administration at the federal, regional, and woreda levels, as well as enhancing the EMA and higher learning institutions that contribute to such capacities; (e) developing public information and communication programs; (f) improving legal recourse for tenure security and dispute resolution; and (g) improving capacity for land valuation. The dual system of formal land rights in urban areas is ineffective, administratively expensive, and operates with fragmented, non-market pricing of land. The existing land tenure system in cities is extremely complicated and resource-draining. First, the “permit� (rent) system still prevails (even in cities where land leasing was already introduced, not more than 4-7 percent of land holdings are leaseholds) and requires enormous resources to maintain it.33 Second, the land lease system, which was envisioned, as explained earlier, as a replacement to the “permit� system, has numerous complexities such as a lack of unified payment schemes, a need to calculate an interest on an unpaid part of lease price for auctioned land, etc. It is not possible to implement this system effectively and efficiently in Ethiopia. In particular, both city administrations and lessees, who bought their leases on auctions, have been swamped by a growing number of conflicts around the payment calculations. Moreover, the land leasing is generally unpopular with both land users and lenders, and there are no voluntary conversions of “permit� tenure into land leases. Third, the land leasing, as it is practiced, still did not improve the overall land pricing. The rent paid by “permit� tenants is set up administratively, at very low levels and does not depend on market value of land. Only a negligible fraction of all existing land holdings have originated through auctions, which provide some approximation to the market values. Most newly-allocated land plots are given to wait-listers for housing construction at administratively established prices (usually the same price for the entire city) that are substantially lower than market prices of these plots (usually, from several to ten and more times lower). Similarly, heavily subsidized land is given to many industrial and social land users. Not surprisingly, some sites obtained through waitlists quickly find their way in the real estate market, thus creating windfall profits. While this situation poses a problem in all cities, its magnitude is not known. Obviously, gaps in land payments for similar land plots created by the existing system of “rent� payment, administratively established lease payments, negotiated and auctioned leases create unfair conditions for various land tenants and permanent distortions to the land market. Overall relevance of such a complex and expensive instrument as leasing seems questionable, especially given Ethiopia’s relatively low literacy level. The issue whether the current complicated and expensive to administer system of urban land tenure is good for country’s economic development needs deserves further frank policy debate in the country. If the leasing system will be retained as a form of land tenure for the foreseeable future,34 efforts should be focused on making substantial improvements. First, 33 For example, with the number of “permit� plots in a range of 27,000 – 33,000 in cities such as Bahir Dar and Mekele, this implies issuing 100 - 125 renewed permits per work day, all year around, year after year, resulting in an obvious waste of resources. Payments for “permit� land are very low and set up administratively, without any connection to market value of land. 34 Two realistic alternatives to making urban land tenure more efficient and equitable can be considered: 34 An alternative would be to the dual land tenure system by a single unified tenure form similar to the “fee simple tenure� operating in New Zealand and other countries with British heritage. This would require a more significant 42 leases should be made more attractive to the public and existing disincentives for leasing should be removed or at least reduced. Second, the leasing system can and should be simplified and standardized, in order to become easier and cheaper to administer. Third, the quality of lease contracts should be improved. Fourth, practices of widely subsidized allocation of vacant land for artificially low lease prices should be urgently revised toward more sustainable approaches that would improve cost recovery by the government. In particular, valuable land should be released only on auctions (“auction only zones�). Finally, release of new land should flexibly reflect demand and be based on monitoring this demand.35 In any case, substantive improvements in how the urban land tenure operates cannot be achieved by the government alone. Direct constructive cooperation with the private sector and non-governmental experts is essential. Furthermore, this cooperation should incorporate local practitioners – both governmental and private - from the cities that already accumulated experience in leasing, such as Bahir Dar and Mekele. In other words, improvements of land leasing or design of another land tenure form should start at the local level first, and only after that be generalized by the federal government. Land allocation in urban Ethiopia is not sustainable spatially, fiscally, and socially. In connection with residential land, this issue represents maybe the biggest urban challenge for the government. On one hand, having a land plot for housing has a high social value in Ethiopia, and urban populations in many cities expect to obtain sizable land plots practically for free. For example, in Gondar, with its population of 210,000, about 15,000 people were on the waitlist for cheap residential plots in 2006. In Bahir Dar, with the waitlist about 10,000, and with the then land consumption standard of 250 m2 per plot, dissolving this waitlist would take 333 hectares. Cities do not have deposits of vacant land of such magnitude, and used to expropriate land from farmers around the city. Expropriating land from rural land holders for plots for urban development creates inequality and has high social costs. Moreover, the financial costs (compensation plus infrastructure provision) are generally not recovered from low lease payments by urban land recipients. For example, estimates for Oromia demonstrated that the existing system does not recover even 30% of the cost, even if future lease payments are counted as capitalized. This problem – that social expectations of the urban population regarding cheap land exceed feasible levels of land supply – is typical for many cities, though specifics (such as size of plots allocated, length of waiting lists, etc.) vary. Solutions require serious revisions of current land and housing policies and practices.36 Moreover, as recently demonstrated by Alain Bertaud,37 even the standard of 75 m2 as a minimum land plot size per household (in year 2002) reform by granting perpetual, transferable, non-contract based land rights to urban land tenants, but would preserve the Constitution provision that the land is owned by the state. 35 Detailed discussion can be found in: (i) Urban Land Administration and Land Markets, ONRS. 2006. - GTZ-IS for the Capacity Building for Decentralized Service Delivery Project / Urban Development Capacity Building Office / Ministry of Federal Affairs / The Federal Democratic Republic of Ethiopia / World Bank; (ii) Urban Development Strategy for ONRS, 2006. GTZ-IS for the Capacity Building for Decentralized Service Delivery Project / Urban Development Capacity Building Office / Ministry of Federal Affairs / The Federal Democratic Republic of Ethiopia / World Bank; and (iii) Kaganova Olga, 2005. Establishing More Efficient and Effective Land Management in the Amhara and Tigray Regions, Ethiopia. - Urban Institute report for the Capacity Building for Decentralized Service Delivery Project / Urban Development Capacity Building Office / Ministry of Federal Affairs / The Federal Democratic Republic of Ethiopia / World Bank. 36 Some recommendations in this regard were provided in the reports referenced in the previous footnote. 37 Eco2 Cities. – World Bank, 2010, pp. 323-324. http://siteresources.worldbank.org/ INTURBANDEVELOPMENT/ Resources/336387-1270074782769/Eco2Cities_synopsis.pdf 43 is also not sustainable, as it is not affordable to the majority of the population and thus pushes them into informal housing. The solution Bertaud suggests, with references in many places in the world where it has been adopted, legally or de facto, is to regulate a size of a developable land plot only, but not a number of households living on it. Land-related revenues In addition to the agricultural income tax, regional states are constitutionally empowered (Proclamation No.131995, Article 97, Sub-articles 2 and 10) to collect and utilize rural land-use fees pursuant to federal Proclamation No 89/1997 (Article7, Sub-article 1 and 2). The valuation for the purpose of rural land use rent/tax on smallholder farmers normally is set in terms of ETB/ha irrespective of location, land quality, or actual annual production. Thus, the valuation is not related to an imputed value of agricultural land, although a somewhat more precise notion of holding size was established through the registration process. The cost of collecting property taxes is minimal due primarily to the simplicity of the tax legislation and the concurrent collection of agricultural income and land use fee. Another factor that tends to further decrease the costs of collection has been landusers’ perception of property tax payment as a proxy for title. This perception has resulted in high rates of self-assessment/declaration among taxpayers. Nonetheless, the contribution of rural land-related taxes to regional government revenues is relatively small. For example, in Oromia, where rural land-use fees comprise close to 60 percent of state revenues derived from farmers, the land tax amounted to a mere 7.2 percent of total state tax revenues in the 2006–09 period. Revenues from land users’ payments of rents and lease fees constitute a significant share of municipal income, exceeding 50 percent in many cities according in the middle of the last decade (World Bank 2007). In many cities, the old rent system applies to much of the landscape (more than 90 percent in 2007). However, the newer lease system, which entails somewhat higher fees, increasingly is producing a larger share of revenues, especially in larger cities. Land revenues from land would have been much higher if the rates charged had been closer to free- market levels (as approximated by the informal secondary market in land). Both land rents (charged under the old permit system) and lease rates are extremely low, and fetch less than 5 percent of the cash market (informal market) value for the same type of land. Rent rates are less than 50 percent of the rates of commercial leases, and even less for residential properties (World Bank 2007, 98–99). Although the newer lease system allows for different types of pricing (administrative rates or closer-to-market rates through auction and negotiation), evidence shows that a few years ago, in localities where lease was already practice, the most leased land was still released at artificially low administrative rates. For example, in the mid- 2000s in Mekele, only 5 percent of lease land was auctioned or negotiated, and in Gondar, less than 8 percent. In Bahir Dar 33 percent of lease land was auctioned or negotiated, and apparently a higher share in Addis Ababa. It should be noted that due to low quality of the auction process, even auction prices were below the market level. The negotiated prices were higher than the extremely low administrative prices, but well below the auction prices. The large differences between market rates and the lease rates actually charged by government create opportunities for rent-seeking and bad governance, because the decisions regarding whether land will be released by auction or otherwise is at the discretion of officials. Compared to common international practices, the manner in which lease fees are charged for negotiated and auctioned land in Ethiopia is unusual, and creates difficulties in managing leases. In most countries, lease fees are charged as a stream of regular payments over 44 the life of the lease, adjusted as necessary for inflation. In Ethiopia, however, the total value of the lease is expected to be paid up front.38 However, the buyer has an option to pay in installments over a specified period (which may be shorter than the life of the lease). These payments would include an assessed interest for the unpaid portion of the total lease price. This arrangement then becomes, in essence, a mortgage loan from government to the lease holder. There is no required uniformity in the schedule of payments for different leases, and the need to calculate interest adds to the complexity and costs of administering and tracking the accounts. This system of payments has a second drawback. In assessing bids on auctioned lease lands, it is difficult to assess in advance the true present value of different proposals, because the payment profiles eventually will differ and have unknown future interest rates. The level of sophistication required to administer such a system often exceeds the capacity of city officials, who are not trained mortgage processors. Finally, this up-front loaded payment system also has potentially dangerous fiscal implications. Analysis of the database on 341 auctioned and negotiated leases in Bahir Dar provides a useful insight on what can be expected: the revenue stream would quickly and dramatically decline soon after land available for auctioning ends (which is unavoidable, given that the urban land is a finite resource).. Such a steep decline of the expected revenue stream from existing leases supports the concern (see Peterson, 2006) that a current high portion of land lease revenues in overall local finance may be not sustainable in medium- and long run. Levies for infrastructure and property tax. Since changes in land use usually imply increased land values, there should be clear and transparent mechanisms such as levies for infrastructure and property tax that enable the public treasury to capture its share of the increased values. In the urban context, the Lease Proclamation No. 272/2002 (Article 12) states that urban land use may be changed only through a permit granted in writing by the appropriate body and provides that the lease period for urban land, performance of payments, and tax rates are to be varied upon the conversion of the land use. However, as noticed earlier, administrating such changes on the contract-by-contract basis is expensive and ineffective. These variations in payments and tax rates seem to be intended to enable the government to capture a share of benefits arising from changes in permitted land use. Concerning mechanisms for capturing benefits arising from infrastructure development, there are no special taxes/levies for infrastructure. However, the level of infrastructure development is considered a factor in determining land-use or land-rent payment rates. Nevertheless, due to lack of a practice of regularly updating valuation benchmarks, the public treasury is not fully capturing the benefits arising from urban infrastructure development. In overall, revenues from land are lower than they could be for a number of reasons. First, the absolute majority of newly allocated plots are leased out at very low administrative prices or at “negotiated� prices that also are well below market. The latter includes highly valuable land allocated for commercial development in prime locations (along main streets / roads, at water front, etc.). Second, the revenue collection system for rents (on “permit� lands) is not very effective. In some cities, financial departments do not have lists of payees; do not have billing system; do not track who paid who didn’t, which results in low collection. This leads to massive payment evasion and significant arrears (over 50 percent arrears for Bahir Dar in 2002 and 2003). 38 China, including Hong Kong, has a similar approach, but very few other countries follow such a procedure. 45 3. RECENT ACCOMPLISHMENTS AND FUTURE CHALLENGES 3.1 RURAL LAND REGISTRATION AND CERTIFICATION PROCESS To implement Ethiopia’s federal rural land policy, the certification of rural land-use rights is a key land administration activity that regional governments are expected to undertake.39 With appropriate supporting policies and an effective and noncorrupt land administration apparatus and court system, the possession of land documents based on an adjudication process that clarifies and demarcates boundaries can instill a sense of confidence in a landholder. The reason is that the risk of challenges to his/her rights is diminished with the possession of an official document that certifies land rights. Even challenges from government are easier to deflect, and the risk of expropriation for eminent domain purposes without compensation is lessened.40 More secure tenure and clarity of land-rights possession strengthen incentives for land-improving investments and efficient land-market transactions. A registry of land properties also can facilitate better and fairer taxation and provide the information required for land-use planning. Ethiopia has a history of rural land redistributions and eminent domain expropriations with inadequate compensation. In its case, the land certification program was expected to significantly enhance farmers sense of confidence in the declared policies restricting redistribution and ensuring due compensation for expropriation. A decision was taken to adopt two levels of certification, starting with first-level certification.41 Implementation started and progressed differently in the various regions. In 1998–99, Tigray was the first region to start a land certification process. The region used simple traditional methods and a process handled by the kebele administration. More than 80 percent of the rural population in the region had received land certificates when the process was interrupted by the war with Eritrea. Other regions followed by learning lessons from Tigray’s early experience. Amhara, particularly, expanded the approach by taking into account the various political, social, cultural, economic and other dimensions of land administration. Amhara began by building specialized land administration institutions from the regional to the kebele level and refined and strengthened the participatory process. In 2002 the region piloted a certification process, and full-scale implementation started in 2003, when Oromiya region followed suit. SNNPR initiated pilots in 2004, and moved to full scale in 2005. Emerging regions such as Benishangul-Gumuz and Gambella are only now in developing their land registration and certification programs, and no actual certification activities have started; Afar (with an already-approved proclamation) and Somali regions are at the stage of fully developing their legislation. In all regions, program implementation utilizes highly decentralized procedures. Following a community (kebele) meeting to explain the program a LAC consisting of 5–7 members 39 Land registration and certification is used by many countries as a tool to enhance tenure security, facilitate land market transactions, and enable better land administration, taxation and land-use planning (Deininger 2003; Deininger and Feder 2009). 40 Rahmato (2007, 15) claims that peasants covered in his survey consider the right to compensation to be the most important benefit emanating from the certification program. 41 This decision was largely driven by desire for expediency, the aim to learn through a phased process, and the recognition of limited initial capacity. The first-level certification provides good evidence of rights based on records from previous allocations and captures information on adjoining landholders and an estimate of the parcel areas, but the parcels are not surveyed or mapped. 46 (according to most regional regulations) are elected by popular vote. The committees usually include traditional leaders and others of high standing in the community. With assistance from local officials, the LACs assume responsibility for key aspects of the implementation of the certification program: identification of individual plots, demarcation, and boundary-marking; measurement of the plots; and recording personal details to be included in the land registry and the certificate. Disputes can arise at each step of the process. The LACs are expected to resolve them or refer them to the kebele office. The composition of the LACs is supposed to include at least one woman, but this provision has often not been enforced. In most kebeles, no women members serve on the LACs.42 Training LAC members before they start their duties is required, but has not been uniformly done. According to Deininger and others (2008)43 more than 40 percent of the LAC members in SNNPR did not have their training in advance. Approximately 67 percent of the LACs received copies of the regional land proclamation (only 50 percent in Oromiya), but very few––10 percent overall––were provided with written instructional materials. In fact, written materials to explain the land laws and the registration and certification procedures to the general population were not distributed en masse in any of the regions. In most cases, even kebele officials had not received explanatory material (although copies of the regional land proclamations had been received in approximately 50 percent the kebeles). However, the high rate of illiteracy in rural Ethiopia diminishes the importance of written land administration materials to inform the rural general population. Consequently, the effectiveness of the information campaign depended on attending community meetings and on the oral presentations made there. It seems that the information campaign, even although only oral, has been quite effective.44 The transparency of the registration process was enhanced by the requirement to register plots in a public process with neighbors present in the field, rather than in an office process in which documents of varying credibility would have to be used as key evidence.45 The intensity of official supervision and support to the LACs varied across regions. In Tigray, supervision was provided by students with backing from technicians. In SNNPR, the kebele rural development officers provided oversight. In Amhara, the process was supervised by the woreda survey team assisted by specifically formed “land administration teams� comprised of students trained in the relevant procedures.46 In Oromiya, LACs operated without strict 42 Rahmato (2007, 16). Deininger and others (2008) report, based on 2006 data, that overall only 20 percent of the LACs had women participation in the LACs. The heavy workload demands on members of the LAC, and the need at times to stay overnight in distant locations may have limited the inclination of woman to undertake such an assignment. 43 Unless otherwise indicated, the quantitative indicators in this section are derived from Deininger and others (2008), which is based on a large 2006 sample survey. 44 Such meetings took place before the start of registration in more than 90% of a sample of communities surveyed in 2006 (the rate was lower, approximately 70%, in Tigray). Among households surveyed that year, 81% indicated that a household member had participated in at least 1 community meeting pertaining to the land registration process, and 78% of the households considered themselves well informed about the process. 45 The extensive information campaign, systemic coverage, and adjudication in the field are principles of best practice for rural land registration. These practices are based on a successful sequence of World-Bank-funded land registration and titling projects in Thailand that have been replicated in many such projects in Asia and Latin America (Deininger 2003; Bowman 2004). 46 More recently, although after the first-phase certification in the region had been largely completed, Amhara authorities also hired permanent land administration agents at the kebele level to perform as supervisors and as a link between the LAC and woreda land offices. Nearly 1,000 agents are in place at the kebele level. In other regions, supervision was less intense. 47 supervision but could call on experts at the woreda level. To ease conflicting time demands on LAC members, registration activities typically were conducted in the off-season. Generally, the methods used by LACs to identify, measure, and mark boundaries of plots were not precise (for example, hand-held global positioning system (GPS) devices were not used). Primarily, ropes and tapes were used. In Amhara, in close to 50 percent of the communities surveyed in 2006, approximate measurements were made through pacing and eyeballing. While enabling the process to be accomplished at low cost, these methods introduce the potential for errors and inaccuracies in future transactions and transfers. Based on case studies in Amhara and SNNPR, Rahmato (2007, 17) asserted, “The measurement of individual plots was the most unsatisfactory part of the certification process.� Once the registration activities in the field are completed, households receive a preliminary registration certificate identifying their holdings. Upon entering the details of the plots and the landholders for an entire kebele into the land registry book, an official certificate is issued with each holder’s photograph47 and space for a map. In Amhara, temporary certificates valid for one year are initially given. This provides sufficient time for any appeal such as from absentee right holder before the preparation and delivery of the final permanent certificate. Building on the experience from Tigray, in which all the work is done at the kebele level, the land registry in other regions is compiled at the woreda level. Certificates are prepared for issuance and/or signature by woreda or kebele officials. On average, in sampled kebeles that had issued certificates by 2006, there was a gap of approximately 16 months between the completion of registration and the actual delivery of certificates to landholders. The delay is attributed in part to bottlenecks at higher levels in designing and printing the documents.48 In none of the regions, except for very few pilot areas, did the certificates include a sketch of the plot. In both sedentary agricultural areas and agro-pastoral and pastoral areas, the registration program only partially identified, marked the boundaries, and recorded communities’ common-property lands. The registration of communal grazing lands has been undertaken as part of the ongoing first-level certification program in the four main regions in Ethiopia. Nonetheless, the boundaries of this land are not recorded, and the communal land generally is registered in the name of the kebele with no formal process for the local authorities to register individual parcels of communal grazing land with specific local groups (Andersen and Dupuy 2009). In comparison to other international experiences, the indicators shown in Table 3.1 suggest a very good overall implementation performance. For example, in the successful titling program in Thailand, approximately 20 percent of the plots could not be awarded certificates (Deininger and others 2008, n. 27). It is noteworthy that, in Ethiopia, the share of cases in which a certificate could not be issued in Amhara was considerably higher than in the other regions (approximately 11 percent). This higher number was due primarily to the fact that significant and much contested land redistribution took place in Amhara not too long prior to the onset of the registration program. The related disputes and disagreements hindered adjudication and left the status of use rights in a relatively larger number of plots with resolution 47 Except in Tigray, where attaching a photograph was not required. 48 Deininger and others (2011, forthcoming). There were, however also cases of very fast delivery of certificates, as reported in Rahmato (2007, 18). 48 still pending. Partly because of this, and to ensure completeness of the process before computerization, Amhara recently embarked on a large-scale re-checking of the activities. Table 3.1 Village-Level Characteristics of Land Registration (%) Item Total Amhara Oromiya SNNPR Tigray Popular election of LAC 92 85 96 87 100 Average size of LAC (members #) 12 14 9 13 21 LAC includes community leaders 69 74 70 70 50 LAC training before program start 88 93 98 57 100 LAC got land proclamation 66 96 50 61 75 LAC got other written materials 10 22 4 9 13 Kebele officials received written materials 7 19 4 0 13 Public meeting held before program start 93 100 92 91 757 Plots measured by rope/tape 78 33 90 96 100 Landholders who could not be registered in the 3 2 4 3 1 field (%) Landholders who could not receive certificates 4 11 4 2 0 (%) Source: Derived from Deininger and others 2008, table 3. There are numerous differences between Ethiopia’s regions in several aspects of registration and certification: a. Format of certificate. The Tigray original certificate is a blue cardstock, and an update certificate is a yellow cardstock. In the other regions, the certificate is a green covered booklet that contains, in addition to personal and land plot information, a listing of rights and obligations. b. Fees. Certificates were distributed gratis in Amhara, whereas in other regions, a fee of ETB 2–5 (less than US$1) was charged. The other regions levied additional small charges for the photographs to be attached to the document. All of these fees are in line with best practice in systemic registration projects. To induce high levels of participation, it is recommended that first-time registration be regarded largely as a public service––with broad public benefits. c. Women’s rights. The Tigray certificate registers the husband’s name only, whereas the other regions register both husband and wife’s names. In cases of polygamy, the certificate is issued in the name of the husband and the main wife. The other wives get their own certificate, with the husbands’ names often added in second place. d. Photograph. In Tigray no photograph is attached to the certificate. In Oromiya, a photograph of the husband is attached. In Amhara and SNNPR, both husband’s and wife’s photographs are attached. Additional wives (under polygamy) have their photographs on their copies of the certificate. e. Location of land registry book. In both Tigray and Amhara, registry books are kept at both woreda and kebele levels. It is not clear which version will be considered the ultimate authority in cases of discrepancies. In SNNPR, the registry book is kept at the kebele office, whereas the woreda office maintains data at the household (not plot) level. In Oromiya, a condensed registry book is kept at the woreda office; the kebele office stores the original data collection forms. This 2-location system implies a cumbersome registry update procedure. 49 Establishing consistent procedures for updating registration and certification upon life events (for example, inheritance, gift, and divorce) has not received adequate attention. There are significant variations in how updating procedures are implemented, and the ambiguities are serious. The registry books generally are available only at the woreda levels. Furthermore, the formats of the registry books in Oromiya, SNNP, and Tigray are such that updating the books is difficult. Only Oromiya has a clear requirement for changes to be registered in columns in the land registry book specially designated for transfers. However, limited space in the book leads to cross-references to fact sheets, which are not an integral part of the registry book. In Tigray, the manner of recording transfers was left to the discretion of kebele offices, with consequent lack of uniformity and great risk of certificates going out of date. Similarly, in Amhara, where land registries are kept at both kebele and woreda levels, recording changes is to be agreed between the two levels. In SNNPR the procedure is unclear as the woreda registry book has not been fully designed. In fact, except for Amhara, the design of the registry book has been judged by experts to be deficient, with insufficient space to update details and to account for all parcels (Orgut 2010e). Inheritance is a key form of transfer in agrarian smallholder economies such as Ethiopia, but the recording procedures are confusing. For instance, Oromiya requires a new certificate to be issued to the heir(s), but the procedure for its issuance and registry is still being developed. The requirements in SNNPR and Tigray are not clear. However, Amhara recently established a system which requires the original certificate to be void and a new certificate to be issued to the heir(s). In all regions, divorces imply splitting the property (equal shares in Tigray and Amhara, and some share in the other regions). However, the regions have no procedure for revising the certificates and for recording the changes in the registries. All regions require registration of rental transactions (although in Tigray only leases of more than 3 years are subject to this rule). However, recording leases is done outside of the land registry book, meaning that the registry will not contain information regarding the actual user of the land. Regions’ procedures for update and maintenance of the certificates and the registry books are not compatible with best practice observed in other developing countries. In the latter, land certificates are to be reissued upon life events with accurate descriptions of plot-splitting, new owner(s)’ details, voiding original certificates, and parallel changes in land registry books. Furthermore, the capacity to deal with the maintenance of land records and certificates is hampered by the inadequate training of the officials who are entrusted with this task at the local level (Rahmato 2007, 18). The deficiencies in Ethiopia’s procedures for recording transfers are deteriorating the extent to which the land registries reflect reality. Furthermore, over time, cases are increasing in which the individuals using the land do not have valid certificates in their names, even although they have legitimate claims to the land rights. The current land registration system in Ethiopia is paper based. Some initiatives have been made to introduce computerization to the system. The Information System for Land Administration (ISLA) implemented in the Amhara region through BoEPLAU with support from SIDA is the only software that has achieved a somewhat large-scale deployment for managing land administration records. As of June 2010, ISLA had been implemented in 45, or approximately 25 percent, of the woredas in Amhara and approximately 753,800 holdings and 3.8 million parcels have been registered in ISLA. In 17 woredas all the records are computerized. However, in many offices the data in ISLA is not being updated due to a range of 50 factors, including the lack of trained staff and differing priorities. Realistically, the majority of Ethiopia’s registration offices will have a paper-based system for a long time to come. Nevertheless, it makes sense to prepare for eventual computerization in all offices. 3.2 I MPLEMENTATION P ROGRESS OF R URAL L AND R EGISTRATION AND C ERTIFICATION The process of certificate issuance is not yet completed, but plot registration in preparation for certification has advanced quite rapidly. By mid-2010, MoA reported that Tigray had provided 99.6 percent of eligible households with first-stage certificates,49 Amhara 86 percent, SNNPR 67 percent, and Oromiya 50 percent. The slightly earlier figures collected by Orgut (2010e) (Table 3.2) provide more details on the progress. The activities in the 4 main regions since 1998 have resulted in the registration and certification of over 7 million landholdings. Table 3.2 Status of Certification in Amhara, Oromiya, SNNP, and Tigray Amhara Oromiya SNNPR Tigray Total Area (sq km) 170,752 350,000 113,323 80,000 714,075 Population (mil) 17.2 28.0 16.0 4.4 65.6 Rural woredas 128 264 134 34 560 Kebeles 3,146 6,419 3,586 695 13,846 Kebeles per woreda 25 24 27 20 25 Rural households (mil) 3.696 5.314 2.848 0.924 12.782 Estimated total parcels (mil) 19.634 20.335 6.277 3.431 49.676 First-level certificates issued (mil) 2.749 1.800 1.900 0.906 7.355 Completion stated (%) 70 42 65 98 N/A Completion calculated (%) 74 34 67 98 58 Second-level certificates issued 1,612 3,000 0 0 4,612 Source: Orgut 2010e, 3. Via its participatory and transparent character, Ethiopia’s first-level certification program also is successful in both reaching the poor as much as the richer, and in alleviating some of the gender bias that exists in rural Ethiopia (Box 3.1). 49 However, this figure does not take account of the many cases of records being out of date due to the fact that certification started more than a decade ago and the recording of changes due to life events has been deficient. 51 Box 3.1 Wealth and Gender Aspects of Implementation of Registration Wealth Aspects Past experiences in other countries suggest that the introduction of land registration and certification can provide an opportunity for the well-connected and politically powerful elites to have better access to the registration processes and services, and to take unfair advantage by formally appropriating to themselves land rights over communal land (Deininger and Feder 2009). However, a 2006 household survey of the Ethiopian program shows a different picture (Deininger and others 2008; see Attachment 3 table 1). Information regarding the program was equally accessible to the rich and the poor. Eighty percent–84 percent of the households in both the richest and the poorest quintiles in the sample attended informational meetings, and practically equal proportions (79 percent and 77 percent of the richest and the poorest households, respectively, considered themselves well informed about the registration process. The incidence of election into the LAC (thus having influence on land rights adjudication) is not much different between rich and poor (12 percent and 10 percent, respectively––not a statistically significant difference), and the award of land documents does not seem to indicate wealth-related biases. While the proportion of households holding any type of registration document differed among regions given the different starting times of the registration program, there were no statistically different proportions of wealthier and poorer households holding some certificate (46 percent of the wealthiest quintile and 61 percent of the poorest). The share of parcels possessed by the richest quintile that are not yet registered (approximately 12 percent) was statistically significantly higher than the share of unregistered plots of the poorest (6 percent)––not a large difference, but dispelling any concern that the registration process favored the rich. Gender Aspects A gender bias can be observed in the implementation of the land registration program (Deininger and others 2008; see Attachment 3 table 1), thus reflecting the lower status of women in rural Ethiopia. Women have a much lower rate of participation in informational meetings. Only 38 percent of female respondents, against 90 percent of male, attended a pre-program land informational meeting. Only 17 percent of women survey respondents voted in the LAC elections, and women’s representation on LACs has been low (20 percent overall). Nevertheless, in households that already had received certificates, women were well aware of the significance of the land certificate. In all regions, more than 80 percent of the women knew where the certificate is kept, and practically all women knew whose name is on the certificate. However, gender bias is evident in the extent to which a wife’s name was recorded on the land certificate. As one would expect, in Tigray, which has no legislation mandating joint registration of husband and wife, the majority of households (71 percent) had only the husband’s name on the certificate. Even in the two other provinces in which both husband and wife’s names are supposed to be recorded on the land certificate, there was a significant incidence of cases in which only the husband’s name was on the certificate (58 percent in Oromiya and 21 percent in SNNPR). In Amhara, a very small share of the households (approximately 9 percent) had the husband’s name only on the land certificate, whereas the majority (79 percent) had both names, as required by the legislation.50 While significant progress has been made, the process of registration and certification is not complete. An additional 5 million holdings are awaiting certification in the 4 “main� regions alone. In some places, the holdings are registered in the registry books, but holding certificates have not been issued to landholders. Moreover, as shown in the previous section, updating lags behind. The deterioration in the quality of the land registry records is most apparent in Tigray’s experience. This region began registration and certification the earliest (1989), and therefore has had more years in which changes in land possession, transfers, and other life events have taken place. A 2005 study of land registration issues in Tigray showed that, in some kebeles, 50 These data contradict the results of a 2005 study covering 33% of the kebeles in Amhara. That survey claimed that 33% of the certificates had been registered under the husband’s name only (Teklu 2005). 52 information already was significantly out of date (Haile and others 2005). Other data from Tigray recorded the possession of land certificates within a panel of farmers interviewed repeatedly over time. The data showed that while, in 2003, approximately 95 percent of the respondent households had land certificates, only 88.8 percent of these respondents had valid certificates, presumably due to unregistered subdivisions, gifting, and inheritance (Holden and others 2010). A 2009 case study of one kebele in Tigray in which certificates had been issued 10 years earlier indicated that only 82 percent of the household held valid certificates to the land they were using, compared to 98 percent coverage originally (Orgut 2010a, 14). Problems with updating registration are not unique to Tigray. Data from a 2006 study of 4 regions showed that, of cases of inheritance that had been formally reported to kebele authorities,51 only 33 percent were followed by a change in registration52 (Deininger and others 2008). A recent study commissioned by USAID reports that a significant proportion (30 percent) of farmers surveyed indicated that they anticipated subdividing their land among their children, even although it was less than the legally mandated minimum holding size (2008, 26). Their intentions imply that these subdivisions either will not be registered at all; or that, upon the death of the parents, the land will be reregistered as a fictitious whole holding, whereas in reality several separate households with separate parcels will be using it. Such fictitious filings will make the registry inaccurate and will hinder future rental and other transactions related to such land. Moreover, the identification and registration of common property land and of household plots has been very limited. A 2006 survey indicates that while in SNNPR approximately 67 percent of the communities had identified their common property areas, the other regions had not accomplished as much. Only 44 percent of the kebeles in Amhara, and approximately 25 percent of the kebeles in Tigray and Oromiya had identified their common property areas. Household plots also were perceived by many LACs to be outside of the registration program, perhaps because property rights for such land traditionally have been quite secure. In Tigray, house plots were not registered at all, and in SNNPR fewer than 40 percent of such plots were registered. In the other regions, the coverage was higher, with slightly over 50 percent of the kebeles in Oromiya and 63 percent of those in Amhara registering their household plots (Deininger and others 2011 forthcoming). 3.3 I MPACT OF THE RURAL LAND CERTIFICATION PROGRAM Ethiopia’s land registration and certification program has been implemented in a relatively short time and at low cost, necessitating simple implementation procedures. Inadequate training also was a problem in many areas. Inaccuracies and less-than-perfect coverage therefore were inevitable. Nonetheless, most observers judge the program to have been successful in achieving the key objective: reducing tenure insecurity without harming vulnerable groups. The sections below detail the evidence on the different aspects of the program’s impact, focusing on landholders’ tenure security perceptions, level of disputes, land market activity, and on-farm investment. 51 These cases likely are to only part of the actual number of inheritance cases, due to under-reporting. 52 The follow-up by change in registration in Tigray and Amhara was higher than the average at 43 percent– 44 percent, and lower in Oromiya and SNNPR at 24 percent–26 percent. 53 The land certification program has improved perceptions on tenure security. The literature on land registration highlights the role of officially issued land documents in providing an efficient means for landholders to verify and protect their land rights from challenges by others, thereby enhancing their perceived tenure security.53 In Ethiopia, some of the risk to private land rights emanates from the government’s ultimate ownership of the land and its ability to interfere with landholders’ use rights through redistribution and expropriation. In such an environment, land registration may enhance landholders’ ability to obtain compensation and equivalent rights elsewhere. Focus group discussions and responses to a questionnaire administered on behalf of USAID indicated that “Farmers were unequivocal in their sense of improved land tenure security, which they predominantly ascribed to the possession of a landholders’ certificate…� (USAID 2008, 24). It was pointed out that government statements pledging limitations on the extent of future redistributions and amended dispute resolution procedures played a role as well, albeit a secondary one. Similar conclusions are derived from a study of a panel of rural landholders in Amhara. Among those without land certificates, the share of respondents expecting a decline in landholding area (due to redistribution, expropriation) increased between 1999 and 2007 from 20 percent to 28 percent, whereas among those who received certificates after 1999, the share declined from 23 percent to 19 percent. A more refined analysis of the same data using a probit model has shown that recipients of land certificates were significantly less likely to fear a possible future change in land ownership (Deininger and others 2011, forthcoming). While 50 percent of the respondents had experienced changes in landholdings in the preceding 5 years, only 14 percent expected changes due to government-mandated redistribution or reallocation in the next 5 years (poorer households were even less concerned than wealthier households, presumably due to the fact that their holdings were already too small to be targeted for land retrieval) (Table 3.3). An overwhelming majority of the respondents in the same sample (73 percent) perceived compensation for land expropriated to be more likely, or much more likely, than in the past. In fact, 86 percent of the respondents expected to receive compensation upon expropriation. However, a statistically significant difference in the expectations of poorer, as compared to wealthier, farmers is observed. Only 83 percent of the poor expect compensation, as compared to 90 percent of the wealthier. This pessimistic outlook may reflect in part the self-perceived lower bargaining and political power of the poor. Similarly, the confidence required to enable the renting out of land has increased (85 percent of respondents). However, the confidence level of the poor is statistically significantly lower than that of the rich (82 percent compared to 88 percent), again presumably due to actual and perceived differences in political power (Deininger and others 2011, forthcoming). 53 See Deininger and Feder 2009. It is emphasized that these positive outcomes are dependent on several key factors, such as good governance. 54 Table 3.3 Tenure Security Perceptions Topic Total Rich Poor Amhara Oromiya SNNPR Tigray Experienced changes in 50 53a 49a 57 54 41 40 landholding in past five years Expect redistribution in next five 14 17 13 18 13 10 24 years Compensation for expropriation 73 77 69** 68 82 73 55 more (or much more) likely now Expect compensation if land 86 90 83** 90 87 79 87 taken Certificate increases confidence 85 88 82* 91 85 77 87 to rent out land Source: Derived from Deininger and others 2008, table 7. Notes: a Richest and poorest quintile, respectively Asterisks indicate the significance of differences between the 2 groups: ** = significant at 1%; * = significant at 5%. Box 3.2 Willingness to Pay for Land Administration Services Insights on the perceived value of the enhanced tenure security can be inferred from analyses of farmers’ declared willingness to pay for certificates regardless of what they actually paid. Data from a large sample of farmers from the 4 “main� regions (Deininger and others 2008, Attachment 3) describes willingness to pay under several hypothetical circumstances. The data indicate that 70 percent of the farmers who did not have certificates at the time of the survey were willing to pay ETB 8–10 to obtain one. It is noteworthy that this amount is lower than the estimated actual cost of registering and certifying a holding, which is approximately ETB 30. Among those who already had a certificate, farmers were willing to pay on average ETB 11.8 to replace a lost certificate, with some significant regional differences ranging from ETB 4.7 in Tigray to ETB 22.2 in Oromiya. Farmers’ expressed willingness-to-pay for land certification will generally tend to be an underestimate of the benefits of certification at this early post-certification period. This is because of the relatively short time span that has elapsed, limiting the farmers’ experience with the higher levels of income enjoyed as a result of the land-improving investments they have made. Richer farmers were willing to pay more than twice the amount reported by poorer farmers. While the difference is not statistically significant, it may be related to a true underlying assessment by the poor that, with lesser means to invest in land improvement, the added security will not translate to income gains that are as high as those to be enjoyed by the rich. This conclusion is supported by the statistically significant difference between the high proportion (97 percent) of rich farmers willing to pay some fee for a certificate (if they did not have one), as compared to the lower proportion of poorer farmers who were willing to pay a fee (86 percent). The data also reveal landholders’ belief that upgraded accuracy of the land certificate, in the form of a sketch or a plot map on the certificate, would enhance tenure security (current certificates do not have such maps, except in few pilot areas). The reason apparently is that sketches or maps reduce the likelihood of boundary disputes in the future. Thus, 90 percent of the farmers professed an interest in adding a map to their certificates, and 93 percent of these would be willing to pay some fee for the upgrade. Among those willing to pay, the average acceptable fee is ETB 12.6, ranging from ETB 5.0 in Tigray to ETB 16.2 in Oromiya. The incidence of land disputes is expected to diminish with land certification. With clearer marking of boundaries and a participatory process of adjudication in the field, most (85 percent) recipients of certificates in the comprehensive study referred to (Deininger and others 2008) perceive the risk of inheritance disputes to diminish once certificates are held. Descriptive 55 information from USAID consultants reported conversations with members of the judiciary indicating that the number of land-related disputes (that constituted roughly 90 percent of all cases brought to the woreda court prior to the certification program) has fallen considerably (USAID 2008, 25). Similarly, 1 case study reports that the average number of land-related cases at the court dropped from an average of 20 a week to at most 2 a week (Adal 2008). Another case study from rural Amhara notes that a great majority of the respondents to a survey reported a decline in land disputes since the completion of certification. Similar views were expressed by a chief judge in a social court and kebele and woreda officials interviewed for the case study (Rahmato 2007, 21). The ability of LACs to handle and resolve land disputes that would otherwise have gone to a kebele-level court is likely to be a factor in the above patterns. The transaction costs entailed in disputes that are resolved by the LACs are lower; therefore, even although these disputes take place, their implications are less severe. Land market activities through rental transactions are relatively frequent. They enable individuals who can earn more income from a tract of land than its present possessor to acquire rights to use the land by offering the possessor a rent higher than his/her current income from the land. Such an exchange benefits both parties to the transaction and benefits the economy as a whole, as the land produces more after the exchange. However, under circumstances of insecure land rights, a possessor of land may be reluctant to rent out land because of concerns that a tenant could take advantage of the murkiness of land rights to claim some permanent rights in the land he is tilling. With land rights clearly documented in an official certificate, the inclination for the current possessor to engage in welfare-increasing rental transactions is enhanced. Evidence from several countries (China, Dominican Republic, Nicaragua, and Vietnam) empirically confirms such impacts (Deininger and Feder 2009). In Ethiopia, a study of a large sample of farmers in Amhara estimated that certification increases the propensity to rent out land by some 13 percentage points and increases the volume of land rented out by 9 percentage points (Deininger and others 2011, forthcoming). An econometric study using data from Tigray concluded that female household heads, who were more insecure in the past due to customary practices that tended to favor extended-family male control over land, had a higher inclination (compared to males) to rent out land after receiving a certificate (Holden and others 2010). Improved tenure security positively affects on-farm investment. With greater tenure security through the possession of land certificates, farmers’ incentives to invest in improving the productivity of land (through structures, fertility-increasing landscaping) are enhanced.54 A 2006 study in Ethiopia confirmed that tenure security perceptions affect on-farm investment in terracing (which considerably enhances land productivity) (Deininger and Jin 2006). A survey conducted on behalf of USAID in 2008 indicated that farmers claimed that they have increased planting perennial crops and have undertaken more soil conservation. Similarly, in a large survey of 4 regions in 2006, approximately 86 percent of the respondents said that possession of a land certificate encourages soil and water conservation, and 88 percent said that it encourages the planting of trees. The responses of poorer farmers have indicated a statistically significant weaker perceived impact than that reported by wealthier farmers. However, the actual magnitudes of the differences are not very large: 8–10 percentage points (table 3.5). Sixty-seven percent of the farmers also opined that land certification will increase soil conservation in 54 This link has been demonstrated in empirical studies in many countries (Deininger and Feder 2009). 56 common property lands, and 76 percent thought that it will reduce private encroachment on such lands. A more rigorous analysis applied econometric techniques to the data from this survey. This analysis showed that certification (statistically) significantly increased the propensity for undertaking land-improving, soil conserving investments such as terracing and bunding and their extent (Deininger and others 2008). Utilizing Amhara panel data, another econometric study indicated that certification increases the propensity to invest in soil and water conservation measures by some 30 percentage points. A simulation of the implied productivity effects of such investments suggests that, under fairly conservative assumptions, the increased output related to such investment effects in the first year after certification would be sufficient to defray the costs of certification, which average US$3.2 per ha (Deininger and others 2011, forthcoming). A study on the impact of land certification in Tigray found evidence of a significant positive impact on investment in stone terracing but not on bunding (both soil conservation measures). The same study found positive and significant effects of certification on farmers’ investment in maintaining and improving soil conservation structures and in tree planting. The positive impacts on soil-conserving investments underlie (at least in part), the increased farm productivity due to certification, estimated by the Tigray study to be as high as 45 percent in some specifications of the analysis (Holden and others 2009).55 Table 3.4 Perceived Impacts of Land Certificates Topic Total Rich Poor Amhara Oromiya SNNPR Tigray Encourages soil and water a 86 88 80a 95 68 88 85 conservation Encourages planting trees 88 90 81 92 92 76 80 Encourages soil conservation on common 66 69 63 77 61 56 83 property land Reduces encroachment on 76 79 75 85 68 72 92 common property land Source: Derived from Deininger and others 2008, table 7. Notes: a Richest and poorest quintile, respectively Asterisks indicate the significance of differences between the 2 groups: ** = significant at 1%; * = significant at 5%. 3.4 U RBAN LAND MANAGEMENT Unlike the case of rural lands, in which systematic and comprehensive land registration has been carried out in all four main regions, there has been no extensive systematic program for registration and certification of urban real estate. No Federal Proclamation for urban lands parallel to the one for rural lands (456/2005) mandates property certification. Different cities have followed their own initiatives and procedures for property registration and administration, as well as their own efforts of computerization. In the absence of a formal, systematic registration of urban properties, first-time registration normally takes place either at the time of provision of land for new holdings or the transfer of ownership for existing holdings. 55 There are no other studies on the impact of certification on income or consumption, presumably because the intervention is relatively recent in all regions other than Tigray, so that the survey data at hand do not yet reflect such effects. However, the empirical literature dealing with land certification impacts in other countries documented increases in output and income. See literature review in Deininger and Feder (2009). 57 According to reports compiled for the Ministry of Federal Affairs in mid-2006 (DHV Consultants 2006a, 38–39, and Appendix H, 7), the proportion of individual urban holdings registered as a percentage of the estimated formally held housing units stood at an estimated 65 percent in Addis Ababa,56 95 percent in Adama, 75 percent in Hawassa, 50 percent in Bahir Dar, and 90 percent in Mekele. Mechanisms for ongoing updating of records also were reported in all of the mentioned municipalities except Hawassa in the SNNP Regional State. Although the practice is by no means consistent across urban areas, municipalities in major towns keep three types of ledger books (registers): transfers, mortgages, and title deeds. However, these files are separate documents and it usually is difficult for any third party to access them. Thus, private encumbrances, if registered at all, are also listed in a separate document from the land registration books. Public restrictions on urban land, especially land-use-planning-related restrictions, are applied in urban areas. Nevertheless, the restrictions on specific properties are generally not part of the record.57 Urban ledger books for title deeds, the nearest thing to a register in some of the urban centers, are stored as files for each property identified by physical address. Urban property records, where such records exist, normally can be accessed only by the holder or his/her legal representative, and commercial banks seeking information on previous mortgages. Everyone else must establish specific interest in the property and secure a letter from the relevant authority, usually the Authentication Office or a court of law, to receive copies or extracts for a fee. The timely provision of access to urban property records varies across urban areas as well as institutions within the same municipality. For instance, in Addis, while authenticated copies of title deeds and transfer contracts are swiftly provided upon request, other urban areas that lack centralized and computerized systems have cumbersome procedures that take significantly more time. Misplacement or even loss of files also is a serious problem in municipalities and semi-urban areas. Urban cadastral maps and the registries are likely highly inaccurate due to the large volume of informal transactions. Where a facility for land rights certification exists, the process usually follows sporadic procedures and is limited to central business districts and to newly allocated land under the lease system. Land that is held under the old “permit� system, or land incorporated into city boundaries from the periurban rural hinterland due to the expansion of cities, is rarely registered (Orgut 2010a, 10). It is in these urban margins that informal settlements are more often located and that tenure insecurity is more prevalent. Similarly, the areas under the old permit systems are those that are more likely to have transferred informally multiple times (due to the longer period permits have been in use), thus possibly subject to challenge due to improper transfer in the past. In countries with fast-growing cities and ever expanding informal urban settlements, such as Peru, large-scale systematic land registration programs have been initiated to regularize the status of these settlements, to enhance tenure security, and to facilitate the incorporation of such settlements in the cities’ infrastructure development plans and services network (Field 2005). 56 All privately held residential properties or an estimated 270,000 of the estimated 420,000 housing units in Addis Ababa have been reported as covered. 57 Exceptions exists under the land lease management of some cities, as discussed below 58 There are no comprehensive data on the extent of informal (squatter) settlements in Ethiopia’s cities, but it is believed, based on the scattered evidence available, that informal housing construction has been a main source of the increase in urban housing stock in recent years. For example, data for Addis Ababa suggest that illegal housing may account for 29 percent of the total housing stock (33 percent of these units are slum housing). Data for Bahir Dar in Amhara for the year 2000 suggest that 26 percent of the city’s housing stock was squatter housing. A study of a sample of squatters in Bahir Dar indicates that they seemed to have a somewhat stable tenure environment. Approximately 20 percent of the sample had lived in the same home for more than 10 years, and 41 percent had been there for more than 4 years. Approximately 50 percent of the sample had purchased their homes from earlier squatters, and most of them had connection to electricity and access to private or communal water taps. However, for the most part, toilets were not available because there was no connection to the sewage system (World Bank 2007, 158). Such a major infrastructure investment apparently is difficult to undertake once settlements have been established in an unplanned and illegal fashion. The informality of “ownership� in these squatter settlements implies that the squatters have no legal claim to compensation when city authorities designate the settlement’s land for more attractive development projects and proceed to expropriate it. This source of tenure insecurity may affect negatively residents’ incentives to improve their housing and to upgrade their neighborhoods from their present slum characteristics.58 The federal legislation on the administration of urban land, that is, the lease proclamation, does not address the formalization of informal holdings. However, there have been some initiatives to formalize existing informal settlements based on requirements set by regional and municipal laws such as the Wuzef Serawoch (backlog cases) project office in Addis Ababa and similar attempts in Dire Dawa and Hawassa. While enabling many informal settlers to formalize their holdings, the legal basis for these initiatives is tenuous at best since the requirements for formality or legality often are unpublished. In addition, the intermittently applied process took 4–5 years and involved hidden costs, making it unaffordable. Moreover, the application of the directives, which was conducted with little consistency and transparency, has been discontinued for unexplained reasons prior to benefiting all eligible informal holders. Generally, Ethiopia’s major cities’ ability to undertake mass urban squatter regularization initiatives (see box 2.5 on Peru) is constrained by the scarcity of professional capacity (particularly among skilled land administration specialists) and the unavailability of modern infrastructure for an efficient urban land registry. The federal and regional offices that process urban land administration matters are similarly constrained by low capacity (Orgut 2010a, 11). These constraints are evident in the very slow pace of formalization and regularization observed in Addis Ababa. In mid-2005, of a total of 200,000 informal and semiformal parcels, registration documents had been produced for only 25,000 parcels and 54,000 cooperative houses. The de facto policy seems to regularize only those informal housing units that were built to an acceptable construction standard, whereas those that are below standard and located on illegally occupied land are likely to be removed, adding to the perceived tenure insecurity in such areas (World Bank 2007, 107). 58 Recent studies in Peru (Field 2005) and Argentina (Galiani and Schargrodsky 2005) show that squatters whose status was regularized (titled) have invested significantly in improving the quality of their housing. 59 The deficiencies of the land expropriation and compensation procedures that were elaborated in this report’s discussion of rural areas and their deleterious effects on tenure security are equally valid in urban areas. If they determine that a better development activity (whether by a public or a private entity) can be undertaken on a parcel of land, urban authorities have practically unlimited ability to expropriate that land (even if inhabited and used for residential or commercial purposes over a length of time). While the main areas subjected to such expropriations in recent years have been periurban rural lands that were needed to accommodate the expanding urban population and economy, properties within long-established urban areas also are expropriated. In fact, the inequities due to inadequate compensation that were highlighted in the rural discussion potentially are more pronounced in an urban context. The reason is the paucity of qualified property value assessors to undertake the statutory valuation of the expropriated property. The alternative proposed by the legislation, namely, the appointment of a committee by the expropriating authority, is handicapped a priori by the lack of independence. Hence, its determination of compensation levels is likely to be biased. There is no higher authority to which the valuation can be contested. This situation was assessed in the World Bank’s 2007 report on urbanization in Ethiopia to result in “a negative and inequitable impact on both the investment climate and on the livelihoods of people� (World Bank 2007, 103). However, promising progress with legalization of urban informal settlers takes place, for instance in Mekele, Addis Ababa, and Dire Dawa. In particular, in Mekele, during the preparation of the new master plan in 2006, about 13,000 informal land holders were identified. The legalization process was completed in 2 years and covered about 12,800 holders.59 Urban Land Management Management of land leases is evolving in many cities. Cities like Bahir Dar and Mekele now maintain separate records of leases and aim to monitor compliance of lease holders with lease contract requirements. For example, Bahir Dar has a database of 439 investment leases. Mekele uses auctions for allocation of most new leases,60 monitors compliance of lessees with contract conditions regarding finishing the construction within 33 months upon obtaining the lease.61 There are also efforts in some large cities to change a strategy of urban growth from further sprawl to redevelopment of their central core parts. Mekele has been actively implementing this approach. First of all, it has a new Structural Development Plan with very flexible land use zoning that encourages high-density redevelopment of the central part. Second, it implements one of its new Local Development Plans in the central kebele, which in fact is a redevelopment plan. It is based on re-parcelization of the area, in order to increase an average land parcel size from the current 37 m2 to one that allows multi-floor construction. Current landholders are not forced to reallocate or do anything as long as they do not try to get building permits. When, however, a landholder wants to build, his choice is either to cooperate with 2-3 neighbors, in order to jointly develop a new parcel planned by the city or sell his parcel to the city and get a replacement site (140 m2) elsewhere plus a compensation for his current improvements. The 59 Based on a special directive approved by the regional Government, the City administration cooperated with the special communities established in each of 7 sub-districts 60 Exeptions are the special policy of subsidizing the lease in the industrial zone and of allocating free land for housing condominiums for the poor. 61 Lessees are given an additional three years to complete construction, while charging fines. If the lease is canceled, the site is auctioned anew. 60 City of Gondar intends to implement a similar scheme of re-parcelization and redevelopment based on public participation and the idea of participatory spatial planning. For example, residents requested establishing an urban park, and the government intends to include this into the redevelopment plan. Some practical steps in various regions have been made to establish constructive linkages between urban and rural land management. Lack of harmonized linkages between urban and rural land rights, and the urbanization strategy that has been expanding cities at the expense of rural landholders is widely recognized, including by the Government at all levels. Comprehensive solutions are not clear yet, but positive examples already exist. These include Tigray’s effort to handle registration and certification of land in towns without a municipal status by rural land administrations. In Amhara, a number of positive steps are also taking place: (i) a joint committee has been established between the agencies responsible for rural and urban land administration to research urban-rural linkages and suggest solutions for peri-urban areas and produced a report with recommendations; (ii) the regional rural land administration agency will assist Bahir Dar to register landholders in the rural kebeles under the city administration; (iii) a new urbanization policy is introduced in Bahir Dar and Gondar, switching from spatial expansion to density increase in city center areas (“going vertical�). 61 4. OPTIONS FOR BUILDING AN EFFECTIVE LAND ADMINISTRATION SYSTEM The preceding chapters demonstrate the significant achievements of Ethiopia in improving its land policy by introducing at both the federal and regional levels detailed legislation and regulations designed to enhance tenure security and equity. In particular, the rapid implementation, at relatively low cost, of a massive rural land certification program was shown to have achieved wide coverage. Land certification was shown to have improved perceptions of tenure security and lead to economic gains in an equitable manner, including advances in the legal standing and actual economic performance of women. The discussion also highlighted areas of deficiency, both in certain aspects of the legislation and in the administrative infrastructure and services intended to implement land policy and to handle land administration. This chapter highlights options for addressing these areas to further develop the land administration system. 4.1 S TRENGTHENING LEGAL AND REGULATORY FRAMEWORK Existing legislation at both federal and regional levels deals inadequately with some aspects of land rights. These shortcomings exist despite the significant progress made in many areas and are reflected in a degree of tenure insecurity that leads to loss of efficiency. Furthermore, the areas of inadequacy likely will cause severe negative livelihood difficulties for the poorer and weaker segments of the population. Adding legislation in areas of urban land and in certain “emerging� regions as well as fine-tuning legal and regulatory measures would lead to gains in both the efficiency and equity objectives. Besides, on some issues such as for reducing a need in land expropriation, change of current policies and practices is needed. The following areas require particular attention: 4.1.1 Strengthening urban land legislation including urban-rural coordination Substantial ambiguity hinders the existing urban land administration and land management rules and procedures and the coordination between rural and urban land legislations. While urban land is not a main focus of this report, however, as demonstrated in Chapter 2 and 3, for sustainable – territorially, economically, and socially - development of cities and towns and sustainable conversion of periurban land in land for development, some policy, regulatory, and practice changes are needed. Most importantly, local urban authorities need to recognize the rights of rural rights holders of land covered by urban plans for conversion to urban land and that these rural rights are converted into urban rights with an equal level of security. Another critical task is to introduce rules that would allow legal holders of land in these areas to benefit from the conversion and resulting increases of land values (for example, by participating in land redevelopment or gaining from it financially). These arrangements should also introduce fiscal instruments (“land development fee,� land dedication for public purposes, etc.) that would capture some part of the increased land values into a public budget and earmark this revenue for investment in public infrastructure. Ethiopia’s economy would also benefit from replacement (gradual but real) of the current dual urban system (“permits� and leases) by a unified system of land rights that would (i) be simpler and easier to administer, (ii) cost less to the government and the land tenants, and (iii) pave the way for increasing government revenues from urban land. Within the Constitution, 62 while retaining the state ownership of land, two feasible options were presented in Chapter 2, section on Land Management. It is recommended that the choice is discussed, made, and the selected policy is codified in required modifications of the federal and regional proclamations, and implemented by local governments, with facilitation as needed, including the transitional period. Besides, some corrections in existing proclamations and regulations at the federal and regional levels are needed, to allocate more decision-making power to local governments on technical issues of land management (land pricing, land lease contracts, land use categories for spatial planning, etc.) Strengthen the legislation on land registration in urban areas and fully ensure recognition of rural land rights. In urban areas, lack of accessible and up-to-date information that can provide local governments with background information for administrating land payments (currently, land rental and lease payments) and facilitate land transactions by private citizens, businesses, and the government itself is a serious shortcoming. Therefore, a special legislation on urban land registration is needed, which would guide and introduce some uniformity in registration and certification. Such legislation would specify: (i) how the system would operate, that is, the norms, standards, and procedures to be established and followed; (ii) which rights and transactions are subject to registration; (iii) access to information. This legislation also needs to address the organizational side of the urban real property registration, but to the extent that allows for flexibility of institutional arrangements. As a final “destination,� conclusion of urban land / property transactions and registration should be separated from local governments. This legislation should also clarify the registration of land rights in peri-urban areas such as rural kebeles covered by urban development plans, in “rural municipalities� and other types of urban centers where the registration is currently hampered by unclear division of responsibilities for registration between rural and urban land administrations. 4.1.2 Eminent domain land take-over and compensation As pointed out, the eminent domain powers of government at all levels are quite extensive. They enable the take-over (expropriation) of land used by private individuals or entities not only for public activities (schools, roads, public utilities) but also for private development considered of superior merit by officials. In some cities, legal tenants can be displaced for none-conformity of their land uses with new structural development plans or master plans. Such wide discretionary power opens opportunities for abuse and graft and causes economic losses to both the private and public sectors. There are two avenues for reducing negative impact of the current practices, and both should be used. First, the need for expropriation should be re-thought and reduced and replaced, to the maximum extent possible, by non-expropriation solutions. Thus, as already mentioned, urban planners should be tasked to reduce the need for displacement and expropriation when they develop new spatial plans. Further, techniques that allow developing or re-developing land through voluntary participation of tenants on the subject territory - the land re-adjustment and re-parceling technique or “land pooling� – should be further tested and applied in Ethiopia.62 Land re-adjustment is an instrument of releasing land for new development, in particular, for delivery of new residential land in the urban peripheries. In land readjustment, negotiation between the city and landholders replaces compulsory expropriation and compensation, and is 62 See a special chapter on this in Eco2 Cities, and a detailed example in Ballaney (2010). 63 leveraged by the increased value of the land after its conversion from farming fields to building plots. Individual land holdings are reduced and the land thus freed is partly reallocated to other occupants who have given up their own land to make way for roads and public infrastructure, and partly sold to outsiders. A critically important element here is that the previous landholders who retained smaller portions of land after re-adjustment should be permitted to freely put this land or its parts on the market. Moreover, new land subdivision plans made especially for the process of readjustment, should contain subdivision for land retained by the old tenants, in case if they want to sell. For “redevelopment� of territories, Mekele’s redevelopment of the central city summarized above in Chapter 3 provides a good illustration of constructive approaches that balances interests of the existing landholders with City’s needs and interests. Second, the power of government to expropriate and define a compensation needs to be subjected to checks and balances through an effective and impartial ex-ante review procedure that includes representatives of civil society and the public. This power also requires an ex ante appeals procedure that, similarly, is representative of the various stakeholders. Current legislation mandates neither. There is some procedure to appeal levels of compensation, but not the eviction decision per se. Since, under current legislation, eviction decisions may be undertaken by authorities at different levels (federal, regional, city/municipal, woreda, kebele), the review and appeal procedures for each of these levels need to be specified. Experience worldwide suggests that appeals procedures should, to the extent possible, not involve the regular justice system, at least in the initial phases of review and appeal. Courts deal with a daunting variety of legal issues. Encumbering them with duties related to eminent domain decisions will both add to the backlog of pending decisions and cause prolonged delays and periods of tenure insecurity while holding up potentially useful projects. In the United Kingdom, the expropriating authority must publish notice of the proposed expropriation that includes full details of the areas involved as well as the map. All affected individuals or entities must receive notice as well. They are entitled to make objections or representations. These must be heard by an inspector from the authority proposing expropriation, that is, an official who is independent of the decisionmakers in the operational department involved in the expropriation. If the objections cannot be resolved, a public inquiry or public hearing is held to address the contestants’ arguments. These procedures are strictly controlled by statutory guidelines and timeframes. A review of eminent domain legislation and practice in Ethiopia made eight recommendations, based on best practices (Anteneh and others 2007, xix and 103–04): a. Expropriation law should (1) establish a clear, fair, and equitable process that protects landholders from unfair or improper actions by limiting the scope for officials to act unilaterally and (2) prescribe implementation procedures that are consistent across the country. b. The law should specify an adequate definition of “public interest� purposes. c. Regional councils should issue directives that clarify the responsibilities of the entities involved when rural land and residents are affected by urban expansion. The directives should ensure that when urban authorities expropriate rural land, the consent of woreda authorities is secured, and actions are coordinated. d. Prior to approval of projects involving expropriation, cost assessment and poverty, social, and environmental impact assessments (EIAs) should be undertaken to guide authorities’ decisions. The cost of the suggested expropriation should be estimated in 64 a detailed and reliable way, and sources of covering these costs identified and secured, before the project can be approved. e. Public participation and transparency should be encouraged throughout the process, particularly at the early stages. Full dissemination of information should take place so that those affected are fully aware of their rights and the appeal procedures available to them. f. The expropriation decision should be open to public challenge; objections should be considered by an independent body before confirming or modifying the proposed expropriation. g. Reasonable and adequate notice of the intention to expropriate and notice of taking possession should be provided to the landholders, and the timetable should accommodate landholders’ basic interests (such as sufficient time to harvest crops). h. Land expropriated for utility infrastructure should follow the normal procedures, with the provision that there should be a fast-track procedure to facilitate faster resolution if timely completion of the public works is essential. Landholders share a widespread sense that, based on actual experiences around the country, the level of compensation for land takeover is not adequate. There are two principles of overall fair compensation that are broadly recognized internationally as good practice: 1) The economic well-being of a landholder should not be harmed by the expropriation and displacement, and 2) A compensation package should not be automatically reduced to monetary compensation alone, and in some cases must include other forms as well. The compensation package should, in general, have several components and depend on specific circumstances. One of the components should be a compensation or in-kind replacement of the “market value�63 of the expropriated land and improvements (house, well, livestock sheds, etc.). For assessing the market value of the expropriated property – as background for defining one of the components of the compensation package – qualification of valuers is important, and massive training of valuation methods applicable in Ethiopia should be provided. Added should also be compensation for the direct costs of resettlement (moving costs) and the compensation for the loss of business (such as local clientele) or movable property. Finally, resettlement assistance should be available if needed. A review of compensation practices in Ethiopia provided a number of useful recommendations (Anteneh and others 2007, xix, 104–05). Many of these recommendations set out measures to make the existing law more effective.64 However, a number of changes in legislation related to compensation were identified: 63 According to the International Valuation Standards Committee (IVSC), the definition of market value is “the estimated amount for which a property should exchange on the date of valuation between a willing buyer and a willing seller in an arms’ length transaction after proper marketing wherein the parties had each acted knowledgeably, prudently and without compulsion.� (www.ivsc.org) 64 The recommendations, which are reflected in the current legislation but not necessarily fully or consistently implemented, include (a) ensuring broad representation on valuation committees; (b) implementing a transparent process to negotiate compensation with the underlying objective of leaving the landholder in the same financial situation as before expropriation; (c) valuation to be based on market prices that recognize the replacement cost of 65 a. Provision for recourse to a specialized land court or tribunal where a negotiated settlement is not achieved. b. Changes to the compensation law to recognize situations where expropriation diminishes the income earning capacity of adjacent land not expropriated. c. Payment of compensation in advance of land take-over. 4.1.3 Legislation concerning rural land rights In the past decade, the federal proclamations concerning rural land tenure and use rights as well as the regional follow-up proclamations introduced significant improvements in tenure security for Ethiopia’s landholders. However, not all regions have a regional follow- up legislation. Moreover, as the preceding sections of this report indicate, tenure security perceptions and the efficiency of land markets and land use controls can be enhanced by providing greater clarity on some issues that can be interpreted ambiguously. Furthermore, some procedures and rules, while well intentioned, introduce uncertainty and may even have adverse effects on efficiency, equity, or both. It, therefore, would be useful to ameliorate these inadequacies through legislative corrections or administrative clarifications and training. Land rentals. The federal proclamation on rural lands allows land rentals so long as they do not displace the original holder (chapter 2.2). While the intention is benevolent (protecting holders from loss of livelihood through farming), the outcome implies inefficiencies in the rental market and loss of productivity. The induced rules within the regional proclamations that limit the extent and duration of rentals, and in some of the regions’ rentals, require registration of the rent transactions. The social interest may be protected through such requirement to register rental transactions with authorities upon a presentation (and deposition of a copy in the record) of a formal contract that specifies duration and other terms. It is advisable to have all rentals of more than a certain duration (such as over three years) recorded. The registration would prevent fraudulent take-over of land rights, protect the interests of the party renting out (as their long- term possession of rights over the land cannot be challenged by the party renting in), and facilitate civil legal action in the case of breach of contract by any party. However, the other limitations on rentals––requirement for official consent, limits on contract duration and on the amount of land to be rented, and option for regional government to change unilaterally agreed rental fees on public lands rented to rural investors––should be removed. Land abandonment. The federal and regional proclamations permit take-over and reallocation by authorities of land that is considered “abandoned.� However, the definition of abandonment is not precise. It may include rental to others, or the operation of low-intensity farming by a landholder who has taken up a nonfarm job. This ambiguity tends to limit rental transactions and hinders labor mobility. In Tigray, recent reallocation of land that has been deemed abandoned on the basis that it has not been used for a period exceeding 2 years has lead to a significant increase in the level of land disputes. It is good practice to clarify explicitly that rental of land, or the maintenance of low-intensity farm operation, are not indicators of abandonment. nonland assets; and (d) establishing an internal supervision mechanism or independent body to oversee and publicly report on expropriation processes and matters. 66 Holding size. The regional rural land proclamations or regulations introduce upper and lower limits on agricultural holding sizes for family farms and distinguished irrigated from rainfed lands. These restrictions serve a purpose during the initial certification and registration process; they do not apply to rental arrangements. Communal land. Communal-grazing, forest, or other land in areas of sedentary agriculture generally is regarded as “kebele� land so usually is registered in the name of the kebele administration. However, the legislation is ambiguous about the tenure rights on these lands. For example, in Amhara, the land proclamation provides for the devolution of rights in communal-grazing land to user groups but guidelines for the consultative process and to support registration are not in place. Pastoral areas. The discussion in chapter 2.1 highlights the absence of policies and guidelines that could be applied by any existing institution toward the management and administration of land resources in pastoral and agropastoral areas. The customary governance of the traditionally communally held water and land resources in these areas is facing major challenges due to the economic pressures and conflicting interests from state and local authorities. There is a necessity for legislation specific to these types of areas that recognizes the communal nature of resources and clarifies the legal status of customary systems.65 Afar region already has formulated and enacted a proclamation on land use and administration that addresses pastoral land issues, and is preparing guidelines for implementation. Federal-level assistance should be provided to other regions to tackle the same challenge. 4.2 O RGANIZATIONAL SET- UP AND ADMINISTRATIVE CAPACITY The description and assessment of the existing environment for land administration focusing on rural land in the four main regions in chapter 2 provide the context for this discussion on organizational set-up and administrative capacity. Before making recommendations on appropriate organizational set-ups and administrative capacity, several key topics need to be considered and decisions made by policymakers on how best to deliver land administration services. Given the Ethiopia’s federated structure of government, these decisions would need to be made by the regions with some over-riding guidance from the federal government. These key decisions are the (a) role of federal government; (b) service delivery model; (3) financial model for delivery of land administration services; and (4) establishment of an appropriate management information system (MIS) and reporting system for all land-related agencies. 4.2.1 Key issues influencing the organizational structure Role of the federal government In designing the land administration infrastructure in Ethiopia, account must be taken of the fact that land is a concurrent competence of the federal government and the regions. According to the constitution, the federal government shall enact laws for the utilization and conservation of land and other natural resources while the regions shall have the powers and functions to administer land and other natural resources in accordance with the federal laws. Thus, land administration offices in the regions must be accountable primarily to their regional 65 Such an approach was applied in the early 1990s in Mexico’s reform of the Ejido system (communal agricultural settlements in which land was used individually but belonged to the community, and water was a community resource). 67 governments, while also being accountable for technical matters to the federal government agencies in charge of overseeing national standards. In a federated country such as Ethiopia, most of the actual implementation of land administration activities takes place at the regional and woreda levels. Therefore, the GoE’s emphasis should be on better coordinating, and eventually unifying, land administration functions and organizations at the regional level, while strengthening technical capacities at the relevant federal agencies. The federal level can play a key role in supporting the regions in (a) capturing lessons and best practices in a way that can readily be adopted in the regions; (b) developing policies and standards that might guide the regions on matters such as parcel identifiers, data structures and standards, and public access to information; (c) preparing and developing standard procedures, workflows, forms and record systems and the associated stationery, manuals, training materials and other information that the regions could adopt and adapt to their requirements; (d) designing and developing software tools and associated manuals and training material that the regions could adopt or adapt to their requirements; and (e) developing standards and policies for the spatial framework for land administration and to support the regions in providing geodetic control, base mapping, and other services. The federal government also could play a more active role in funding or arranging the financing of the regions’ efforts to improve land administration. In India, in which land administration is a State (second-tier government) rather than federal responsibility, the federal government has been funding the regions in efforts to computerize land records systems and apply new technology. The proposal for LALUDEP (MoARD 2010) provides a framework for federal support for the regions in developing their land administration systems. Service delivery model The benefits of decentralization of land administration services must be balanced against the costs of providing the services at the local level. Land administration typically is provided by government as a decentralized service. The major rationales for decentralization are access by users, particularly the public, to land administration services and support to the information needs of local authorities. On the other hand, decentralization can increase delivery costs and inefficiency through lack of consistency across geographic areas. There are both the direct costs in providing the service locally as well as the indirect costs of services associated with decentralization such as providing the necessary oversight of the local offices, arranging for the backup of the data held in the local offices, and facilitating access to the records being maintained at the local level. The low number of transactions in a kebele implies that many kebele offices will not be able to support a full-time person. Services indeed should be accessible at the local level. Amhara, for example, already has made a decision to decentralize rural land administration services to the kebele level, However, the concerns of cost-efficiency should be taken into account. Moreover, the facilities available in the kebeles are limited. Few kebeles have been able to establish and maintain registry books. As a result, updates to land records must be effected at the woreda level. In practice, there is little updating of records at the woreda level due to a range of factors from difficulty in access, lack of public awareness of the requirements and benefits of registering changes in rights to inefficient procedures, poor records systems, and a lack of emphasis on the importance of keeping the records up to date. International references to the degree of decentralizing services provision are presented in Box 4.1. 68 Box 4.1 Degree of Decentralization in Land Administration Services in Thailand and India Thailand Thailand has perhaps one of the most efficient manual land registration systems in the world. Its average time to record a transfer is 2.5 hours. The country has area of approximately 513,000 sq km and a 2010 estimated population of 66 million. Administratively, Thailand has 76 provinces including Bangkok, and 878 rural districts and 50 districts in Bangkok. The Department of Lands (DOL) in the Ministry of Interior is responsible for recording rights to nonforest land (approximately 47 percent of the country) under the Land Code of 1954. The department does this through its 76 provincial land offices, 272 branch provincial land offices, and 758 district land offices. The title register is maintained at the provincial level (in either the provincial or branch provincial land offices). The records for lesser documents are maintained at the district level. To support this network of land offices, there are a limited number of office typologies, with standards for each type of office and standards for staffing and equipment of each type of office. There are clear criteria for establishing new branch provincial land offices based on a number of factors. They include the number of titles, projected levels of annual registered transactions, revenue expected to be collected in office, distance people need to travel, and availability of funds. The Thai network was not built from scratch. However, since 1901, when the Department was established, the network has gradually expanded from Bangkok to the rural cities and then into the rural areas as the coverage of the title register has gradually expanded. India In India, which has a tradition of village-based manual records systems in rural areas, the states/provinces have had to make decisions on how these records were computerized. In most states in India that have computerized land records systems, the kiosks to provide access to the computerized records were established at a subdistrict level (taluk/tehsile/mandal). This decision made the computerization effort manageable, but meant that villagers who needed to update records or obtain extracts from the records for various purposes had to travel to the kiosk rather than chase the village accountant who traditionally had maintained the manual village land records. Service delivery at the local level must be cost-efficient. As noted above, under the systems established in the four main regions in Ethiopia to record rural landholdings, kebeles are required by law to maintain the original land registry book. Yet, the reality is that few of the 13,846 kebele land offices operate as intended. However, there are efforts in Amhara and Tigray regions to move responsibility for operational registration from woredas to kebeles. Considerable effort is required to develop the necessary human resources and capacity to deliver a land administration service at the kebele level. Recent studies have estimated the staff training requirements under this model at 14,446 university graduates and 24,081 college graduates (Orgut 2010a, 58). The costs to develop the necessary woreda and kebele offices are estimated at up to US$70 million (Orgut 2010f, 48–49). A simpler and cheaper way of improving the registration procedures is for kebele offices to act as a “window� for the land administration system with the registration to be undertaken at woreda level. Under this approach, each registration application would be submitted and recorded in an application book in the kebele. The approved application documents then would be transferred to the woreda office for processing. Completed registration documents would be returned to the kebele office. This “kebele window� approach could be an interim arrangement until it is decided that the establishment of a fully functioning kebele office is justified. It would be established based on criteria similar to those applying in Thailand, such as the number of 69 documents, expected level of transactions, distance people need to travel, and availability of funds. This approach will require some investment in the process for transferring records between the kebele and woreda offices. However, with 560 woreda offices in the four regions, the effort to build capacity under this alternative delivery model is much more manageable and would bring services much closer to citizens. One of the strengths of the current system is the use of elected volunteers to the LAC at the kebele level. The LACs initially were established to support the certification programs. However, recent land proclamations set out ambitious goals for the future activity of these committees in subject areas such as land conflict mediation, formalization of land rental markets, land-use planning, and monitoring sustainable land use (Holden and Tefera 2008). In the recent business process re-engineering (BPR) in Amhara, approval was given to hire one land official in each kebele. By mid-2010, Amhara had filled 920 of the 3,146 approved kebele-level land administration positions (Table 4.2).66 However there is little evidence that this level of support is necessary, particularly once the initial effort to complete certification has ended. A recent review in Fagita Lekoma woreda in Amhara noted that a total of 575 transactions were recorded by the woreda in the 6 months prior to February 2010. This equates to approximately 3.2 transactions per day for the woreda, or approximately 1 transaction every 8 days in each of the 25 kebele in this woreda (Orgut 2010e, 43). Based on this level of activity, while there is a need to support the LACs in the kebele, it seems reasonable to consider one official as being able to support several kebeles. The ongoing and expanded role for the voluntary LACs also will need to be supported with training and materials, the extent of which is as yet undetermined. Financial model A land administration system can generate significantly more revenue for government than the costs required to fund the various land sector agencies.67 Studies also have demonstrated that users are prepared to pay for good land administration services. A study in Ethiopia has demonstrated users’ willingness to pay for a range of land administration services, including the addition of maps to the book of holdings (chapter 3 above). The ability to recover costs from users and the willingness of users to pay for services has underpinned the strategy of various governments that have implemented major programs to improve land administration systems. Countries implementing major land administration programs with a stated aim that the land administration agencies be self-financing include Albania, El Salvador, Kyrgyzstan, Lesotho, Macedonia, and Romania. Establishing a land administration system that is self-financing ensures that the system is not at risk in future budget cutbacks but is sustainable.68 However, numerous factors need to be considered. First, is it unclear how a self-financing system can be implemented in the context of Ethiopia’s administrative system. In it, substantial responsibilities lie with woredas and 66 The latest figure provided by Amhara EPLAUA is 1,589. 67 In investigating land administration systems based on data compiled in approximately 2001, Burns (2007) found that the ratio of revenue to expenditure for the land registration system was 20.7 for the State of Karnataka in India (staff 1,546), 5.1 in Thailand (staff approximately 8,500), 2.4 in the Philippines (staff 2,408), and 1.6 in Latvia (staff 160). The ratio of revenue to expenditure in most developed systems was approximately 1.0 as the registries operated under a system that set fees at rates expected to cover the cost of provision of services. 68 Other advantages include more flexibility in the use of funds, since the government can free some of the typical government regulations on expenditure with appropriate checks and balances. Self-financing also can focus attention on efficiency and create disincentives to build capacity just for the sake of it. 70 municipalities. The national system lacks regional autonomous agencies that collect all the revenue and hire all the land administration officials in their woreda (and municipality) offices. The implementation of cost-recovery or self-financing needs to recognize that there will be some cross-subsidization because some services will be more in demand than others and some offices will be busier than others. Typically, the offices servicing certain districts of the urban sector, where there usually is a more active land market, are busier, and revenues collected in these busier offices underwrites providing services in rural areas. Second, a system that is self-financing over the long run generally still requires the cost of the first registration activity, or the initial certification activity, to be underwritten by the government. First registration usually is undertaken on systematic and cost-effective basis. Charging the full cost of the activity to landholders can be a strong disincentive to participation, particularly to those less well off. The Thai government charges only the cost of corner marks placed during systematic registration (approximately US$3, or approximately 10 percent of the cost of producing each title). The government more than recovers this initial cost in fees and taxes collected for subsequent registrations. This approach would appear to be the one adopted by the GoE. The draft proposal for LALUDEP includes the proposal that, to encourage participation, the cost to landholders for second-level certification should be minimal. On the other hand, the draft then proposes that, once the land registery is established, cost recovery schemes can be instituted. Third, Ethiopia appears to have no adequate policy for charging fees for land administration services. Amhara and Tigray charge no fees for first-level certification. Other regions charge fees based on recovering the cost of preparing the green book and registry books. Fees also generally are not charged for the registration of subsequent transactions. Only in Tigray is a fee (ETB 3) charged per transaction (Orgut 2010b, 68). The self-funding policy should not be applied office by office, but regionally. This arrangement will mean that the policy will need to build in a cross-subsidy in the fees raised in busy offices to cover the costs of less busy offices. Cross-subsidization will present a challenge in Ethiopia, in which the land administration systems for the urban and rural sectors are separate. Even in Oromiya, which has a single regional entity responsible for both urban and rural land, the main implementation and revenue-collection is done by the woredas and municipalities. Management Information System (MIS) The Management Information System (MIS) needs to be significantly strengthened. An effective MIS for land administration should provide information that enables managers to run their organizations more efficiently and effectively, facilitates better services to be provided to citizens, and enables the government to assess progress on implementing policies and programs. MIS can be used to monitor a range of performance factors relating to finance, service delivery, production, backlogs, human resources, and disputes. The land administration systems in Ethiopia’s four “main� regions (Amhara, Oromiya, SNNPR, and Tigray) employ a range of performance monitoring, varying widely in level of detail and factors monitored. A recent review of the existing MIS procedures recommended that a balance should be struck between having a comprehensive set of statistics and the need to have an MIS system that is manageable in terms of recording, analyzing, and reporting the information (Orgut 2010f).69 69 Orgut 2010f recommends the following key performance indicators to be monitored: a. Extent of first-phase-level registration of holdings 71 4.2.2 Structure of land administration institutions at regional and woreda levels The discussion above indicates a focus on the regional office, which provides operational oversight and key technical support in certification, IT, and survey/mapping to providing land administration services that are provided initially at the woreda level. The kebele LACs provide a “window� or “front-office� that accepts and approves applications to update land records. Land officials would be appointed to support the LACs in a number of kebeles, and systems would be implemented to transfer records/applications between kebele and woreda offices. The cost of certification would be underwritten by government, but a policy of self-financing would underpin the gradual expansion of services beyond the woreda offices. The implementation of a self-financing policy would necessitate an increased status for land administration agencies and some sort of integration between rural and urban land administration systems in the regions. Initial steps to implement integration between rural and urban land administrations already have been made by regional governments. Thus, Amhara has a single Bureau of Environmental Protection, Land Administration and Use that reports directly to the regional government rather than to a sectoral bureau. Similarly, in 2009 Oromiya established a single bureau (Proclamation 147/2009) “to administer and regulate the urban and rural land and preparation of land-use planning of the region in an organized manner under one executive organ.�70 The merger of the hitherto separate urban and rural land administrations in Oromiya has not been straightforward, and difficulties persist. Over the last few years alone, the institutional set-up in SNNPR saw several significant changes, resulting in an unstable situation. Both Benishangul-Gumuz and Gambella regions very recently established separate institutions mandated with land administration. In Tigray, rural land administration nominally comes under the Bureau of Agriculture and Rural Development. Yet, the former is de facto independent, as it has an earmarked separate budget. However, a proposal to grant the land administration team a formal independent status and to merge urban and rural administration functions under one roof has not been approved, apparently due to budgetary reasons. The merits of unified land administration at the regional level are obvious. Hence, in the medium term, it is important, first, to accomplish integration in all regions. Second, the reforms at the regional level should be complemented with appropriate steps at the woreda and town levels. In the near term, improving coordination between urban and rural land administration, especially in periurban areas, is essential. b. Backlog of first-phase-level certification in relation to registration c. Progress of second-level certification in surveying parcels and registering holdings d. Number of male- and female-headed households covered e. Number of registered transactions (gift, inheritance, exchange, rent) f. Number of land related disputes that exist and how many of these go to court g. Level and type of skills that staff have and how these change over time h. Area of investment land made available. A broader set of information, especially for managing urban-rural linkages, might include: a. Urban centers with current Master Plans or Structural Development Plans b. Total land area added by these plans to territories within administrative borders of these cities and urban centers (i.e. the area planned for urban expansion) c. Number of households whose land holdings were expropriated (by rural and urban kebeles separately) d. Number of informal landholders whose holdings were legalized / formalized. 70 Megeleta Oromiya 2009. 72 Specifically, the functions of a regional land administration agency should include: a. Developing and implementing the land administration program71 b. Harmonizing and aligning systems and procedures in the region and cities c. Developing a long- and short-term business plan for land administration d. In consultations with federal agencies, recommending to regional and city authorities adjustments in the legal framework to accommodate local circumstances e. Coordinating land administration in rural and urban areas f. Developing the human and material resources at each administrative level within the region and the cities g. Ensuring consistency of approach and quality of service provision h. Overseeing planning and budgeting and providing human and material resources at each level in the region i. Developing job descriptions for land administration staff at all levels. The field operations of the regional land agency would be handled primarily by offices at the municipal and woreda levels. Based on the experience in Amhara, the initial focus will be to deploy efficient manual land records systems. The manual system initially will be based on the existing first-level certification system without a strong spatial framework. Computerized systems also would be developed and deployed under a clear strategy that would ensure that computerization would be implemented only in offices that were ready for it, in other words, able to operate and maintain the computerized system. As a strategy is agreed for the second-level certification, surveying and mapping will be implemented in the woreda offices. Because undertaking second-level certification is a one-time task that requires specialist skills, the task is likely to be implemented by the private sector under contracts arranged by the regional office. This is the approach that has been put forward in LALUDEP. The woreda office will have some involvement in the oversight of the second-level certification and will take on the task of maintaining the survey and mapping records. Responsibilities of the woreda land office would be to implement the land administration system, including registration and, once first-level certification is complete, surveying plots; ensuring consistency in approach and quality of service provision; and communicating and disseminating information. Note that it is not envisaged that land valuation will be handled at the woreda level, except in large urban centers. 4.2.3 Building up the human resources and physical capacity for land administration The staff available for land administration in the four regions are substantially fewer than the total number of approved positions (Table 4.1). The number of approved positions (including staff for environmental protection) is based on the recent BPR, not on a detailed analysis of the number of staff required to support an appropriate land administration system. Nevertheless, clearly there are shortages of staff, particularly in Oromiya, SNNPR, and Tigray. 71 The proposed functions in this section are adapted (and modified as deemed necessary) from proposals made in Orgut 2010a. 73 Table 4.1 Approved Staff under the BPR and Existing EPLAUA-Type Staffing the Rural Areas of Four Regions Region Region Zone Woreda Kebele Approved Existing Approved Existing Approved Existing Approved Existing Amhara 50 44 124 93 1536 760 3146 920 Oromiya 79 32 180 90 7656 1056 6419 N/A SNNPR N/A 6 N/A 13 N/A 26 N/A n.a Tigray 86 54 N/A N/A 422 181 695 0 Source: Orgut 2010a, 46, drawing on EPLAUA offices in the 4 regions, including land administration and environmental protection staff only. Capacity constraints vary considerably by region. The strategy outlined in the preceding sections of this report provide a basis for estimating the human resources required to provide the initial scale-up of land administration services in the four main regions. The focus of the scale-up will be to build up the regional offices and provide registration service at the woreda levels. The kebele offices will act as “windows� to the registration system. Based on the following assumptions, table 4.7 sets out the minimum “required� number of staff to support land administration in the 4 regions: a. The existing regional staff in Amhara, Oromiya, and Tigray can support the regional offices. SNNPR, which has a similar population and size to Amhara, will require a regional staff of approximately 44. b. The existing zonal staff in Oromiya and Amhara are sufficient. SNNPR will require a zonal staff of 93, the same as Amhara. Tigray has no zones so will require no zonal staff. c. Sufficient staff are available to support the current and any immediate improved system of land administration in urban areas. d. The woreda offices can support an ongoing operation with an average staff of 5 in each woreda office and with additional staff supporting the kebele LACs based on 1 staff member supporting 3 kebeles. e. The additional staff required for major tasks, such as first- and second-level certification and the computerization of records, are sourced under contract over and above the staff listed above, who are responsible for the ongoing operation of the land administration system. Based on these assumptions, Table 4.2 sets out the number of required positions for the existing staff levels. Table 4.2 Required and Existing Land Administration, Land Use, and Environmental Protection Staff Region Region Zone Woreda Kebele Requ. Exist. Requ. Exist. Requ. Exist. Requ. Exist. Amhara 44 44 93 93 640 760 1049 920 Oromiya 32 32 90 90 1320 1056 2140 N/A SNNPR 44 6 93 13 670 26 1196 N/A Tigray 54 54 N/A N/A 170 181 232 0 Sources: Calculated based on the assumptions listed above, plus the information on available staff from Table 3.2 and the number of kebele from Table 4.1. The analysis indicates the severe shortage of staff, especially in SNNPR, even under the “minimalist� model. Assuming that 33 percent of staff need university educations, and the rest require college educations, the following indicative training requirements are identified to 74 support the above staffing levels. The staffing requirements set out in Table 4.3 likely are to be the minimum requirements, because the staff necessary in the sector to undertake the major certification and computerization to build the records systems have not been included. Moreover, the human resource need for land administration in the urban sector is unclear. However, the numbers in the table seem to be a realistic estimate of the minimal requirements for rural land administration, particularly given the existing capacity of educational institutions in Ethiopia. Table 4.3 Anticipated Staff Recruitment under the “Minimalist� Model Region Required Existing New recruits Total University College Amhara 1826 1817 9 3 6 Oromiya 3582 2462 1120 373 747 SNNPR 2003 45 1958 653 1305 Tigray 455 235 220 73 147 Total 7866 4559 3307 1102 2205 Source: Calculated based on the information in Table 4.2 and the assumption that 33% of new staff will require university qualifications. The results of the analysis reflect a substantial need to increase the training capacity, including higher-learning, vocational, and short-term training. Currently, only the Institute of Land Administration (ILA) at Bahir Dar University is offering a 4-year BSc in land administration. (The program started only in 2006, with the first 42 graduates in 2010.) New degree programs require the recruitment of qualified post-graduate teaching faculty, and recruiting high-quality faculty will take time. (MSc graduates from ILA are not expected before 2015.) Additional options will need to be developed at other universities, including those in Mekele, Hawassa, and Adama. These institutions of higher learning also need to develop shorter courses in specialized subjects that target existing and newly recruited land administration staff. Additional practical short-term courses in surveying and GIS that target land surveyors and land registration clerks should be developed in regional colleges in cooperation with land administration authorities and ILA. The physical/infrastructure capacity of the land administration system requires strengthening. Storage facilities for supporting documents generally are poor in woreda and kebele offices. The paper land records often are kept in inadequate rooms with little protection against humidity and rodents. Table 4.9 shows a recent assessment of the environmental constraints for land administration in the regions. 75 Table 4.4 Physical Environment for Land Administration in the Regions Region Zone Woreda Kebele Human capacity Tends to have more Some qualified staff Very few trained Staff with little qualified people and qualified staff training accessible Office Usually dedicated Often based at same Limited Very poor accommodation facilities/office location as a woreda building available Storage facilities Usually available Some potential Some potential Usually none Power Power cuts Power cuts Power cuts Power cuts Telecommunications Tel., Woreda net Tel. Woreda net None Environment No temp./humidity No temp./humidity No temp./humidity No temp./humidity control control control control Security/backup Some secure access Limited secure Locked room, no No facilities access other facilities Fire/disaster (paper No back-up No backup No backup No backup copies) Source: Orgut 2010c, 26. Transportation and other equipment require substantial investments. The above strategy–– to concentrate on providing land administration services at the woreda level, with assistance to land offices that support the operations of the voluntary LACs in the kebeles––gives a basis for estimating the immediate cost of improving the physical resources in the four main regions. The major cost in the estimate is that of the motorcycles for the staff responsible for supporting the kebele offices. If enough staff are recruited for 1 staff member to support 3 kebeles and each has a motorcycle, over 4,600 motorcycles would need to be procured at a budgeted cost of US$11.5 million. Another major cost element in building the physical resources is the likely cost of civil works. There is no inventory of the existing offices so the cost to set up the 560 woredas ranges from US$14.9 million, where no new offices are required, to US$18.8 million, where new offices are required for all 560 woredas (Table 4.5). Note that this initial estimate is based on a minimum investment to establish the 560 woreda offices that will have the ability to service the kebele offices. This estimate does not include stationary and operating expenses. Table 4.5 Indicative Cost of Base Physical Resources to Support 560 Woreda Offices Unit Item # Cost Comment price Civil Works • Office (4m x 4m) 3500 2 x 560 3,920,000 2 offices in 560 woredas Vehicles • Motor bikes 2500 4616 11,540,000 1 for 3 kebele (total 13,846 kebele) Furniture and Equipment • Shelves 80 5600 448,000 10 lots of 1m x 4/office • Lockable cupboard 400 560 224,000 • Computer 1200 1120 1,344,000 2 per office • Printer 800 560 448,000 1 per office • Desk/chairs 100 3360 336,000 6 per office • Field equipment 1000 560 560,000 Minor survey and drafting equipment Total (without civil works) 14,900,000 Total (with civil works) 18,820,000 Source: Updated from the information in the table in Orgut 2010f, 51. 76 4.2.4 Institutional set-up and administrative capacity for urban land administration Revamped urban registration system should improve the existing system rather than be a new system. This choice was solidly agreed by the extensive stakeholder consultations conducted by a review team of experts on behalf of the then-MoWUD (DHV Consultants 2006b). The consultations also stressed that property records should be kept at the local level and administered by a new municipal organization or department. Generally, all stakeholders supported the concept that such an organization should be autonomous and, in the long term, financially self sustaining. Similarly, all stakeholders concurred that registration should start on a sporadic basis but that systematic adjudication should soon follow. A majority view held that the registration system should use a general index map to be compiled by the most appropriate and cost-effective means, that is, using a cadastral survey if one is already available, or remotely sensed imagery if it is not. The most disparate points of view were on the question of a single agency for all cadastre functions or separate agencies for real property registration and valuation/land use. Based on the review, the consultants produced a comprehensive change agenda that can serves as a foundation for a strategy to advance urban land administration in Ethiopia. The two key ingredients of such a strategy are: a. An appropriate institutional framework and organizational structures need to be developed. It is recommended that the institutional structure at the municipal level separate the legal from the fiscal cadastre to avoid possible conflicts of interest, both from the point of view of property owners and from the management of property and land information. A legal cadastre is in the interest of individual owners of rights on real property, whereas a fiscal cadastre is in the interest of the governmental authorities. Ideally, the land agencies at the municipal level should be financially self-sustaining, autonomous registration agencies responsible to municipal governments. These should be agencies dedicated to these specific mandates with a clear customer service delivery mandate and separate from more municipally oriented functions of property valuation, taxation, and land use control. The focus at the local level should be on real property registration. The national and regional levels will emphasize access to data and integrating parcel-based property information with other land-related information based on topographic maps b. The national urban land administration system is to consist of three layers, with agencies and offices to report to their respective federal, regional, and municipal governments. In this model, the national policy, standards, work processes, and ICT strategy will be set on the federal level, whereas the operations are within the municipal Real Property Registration Offices. The Regional Real Property Registration and Information Agencies will provide support to the municipal registration offices and act as regional spatial data infrastructure hubs. There are weaknesses in this arrangement (for example, dual reporting). However, it best fits the Ethiopian policy for regionalization/ decentralization while promoting national standards. It should be noted that, for federal and regional offices to perform their assigned roles, significant capacity building will be required. As demonstrated in previous chapters, a number of issues related to peri-urban areas cannot be addressed by neither urban not rural lines of administration alone and require inter-agency cooperation. International experiences in land management have demonstrated 77 that establishing a temporary cross-agency task force or committee with a clear mandate to develop realistic solutions for specific problems can be a very effective instrument. This coordination mechanism should have commitment and support from the high-level policy- and decision-makers, and work with a realistic schedule and with clear tasks. At the federal level, it can be set up jointly by the MoA and MoUDC with participation of other agencies. This model might then be adopted at the regional level, possibly with membership from the most advanced city-level administrations. These technical groups should be tasked with outlining a concept and implementation procedure for addressing the problems on urban-rural fringes; discussing them with wider stakeholders and the public; establish “prototypes� in a small number of localities; and, possibly, facilitate them for codification by senior policy-makers in new regulation for broad implementation. 4.3 P ROVIDING EFFECTIVE LAND ADMINISTRATION SERVICES The land certification program undertaken in Ethiopia in the past decade has accomplished much, at a relatively low cost (chapter 3). Nonetheless, several issues have come up during this process and the registration system more generally. 4.3.1 Strategic approach to strengthening land administration A significant first step for a strategic approach to strengthen land administration has been made for rural land and should be expanded to include urban land. Specifically, early in 2010, MoARD released a draft design of the (national) Land Administration and Land Use Development Project (LALUDEP). This project with an indicative budget of US$253.7 million was initially designed with 8 components: a. Improving the Policy and Legal Framework for Land Administration and Land Use b. Reforming and Strengthening the Institutional Framework for Land Administration and Land Use c. Conducting Second-Level Rural Land Certification d. Promoting Public Information and Awareness for Land Administration and Land Use e. Developing and Implementing Land-Use Planning f. Developing Human Resources in Land Administration and Land Use g. Developing Research in Land Administration and Use h. Ensuring Project Management, Monitoring and Evaluation. The major cost item in this indicative budget is the second-level certification (approximately 48.1 percent of the projected budget). The design of this activity recognizes that additional work is required to develop and agree on an appropriate methodology to undertake first-level certification. The design then proposes that a major program be implemented to undertake second-level certification for approximately 10 million landholdings comprising approximately 40 million land parcels in the 4 main regions. It is proposed that the certification be undertaken on a systematic kebele-by-kebele basis using private sector firms or contractual staff. The budget works out at a unit cost of approximately US$3 per parcel which is very low compared to international experiences (section 4.3.2, particularly the information in Table 4.7). The other 2 major cost items in the budget are the reform and strengthening institutions (21.1 percent) and the public information and awareness (24.0 percent) activities. 78 While the concept as described above requires some revisions, the Concept Note for LALUDEP provides a basis for developing a program to improve land administration in Ethiopia. It is likely that without such national program, strategic direction and overall progress will be limited. In designing a project to strengthen land administration in Ethiopia it is important to recognize the federal structure and the fact that the requirements of each region differ. The four main regions––Amhara, Oromiya, SNNP, and Tigray––have issued proclamations and implemented certification programs. The status of land administration in the four regions differs. Perhaps the most advanced, Amhara has benefited from significant assistance from SIDA. Afar and Benishangul-Gumez have issued proclamations but have not commenced activity. Benishangul-Gumez has 4 staff in the land sector, so the activity there is building from a low base. 4.3.2 Improved land administration The only way to ensure maintenance of land information at the local level is to have a low- cost and participatory process whereby land records are maintained by the LACs with community input. This process would have implications for the technology chosen and the procedures to be applied, especially the management of spatial data. The process also would require training to build capacity. Other countries (for example, Rwanda, which has chosen such a model) illustrate that such process is feasible. Ethiopia has significant problems with the reliability of data in the land records stemming from the first-level certification. Thus, to strengthen the maintenance of the system in Ethiopia, a number of aspects should be considered.72 One reason for the deficiency relates to the storage of data. Amhara made an effort to record data digitally in 45 woredas, but this goal was not sustained because follow-up transactions were not digitally recorded. In general in the country, information in rural areas is held in paper form, mainly in unsuitable storage facilities lacking climate control, security from unauthorized access, and protection from natural disasters and pests. The availability of computers and advanced mapping equipment is inadequate even in regional land offices (except Amhara’s regional office), and is almost nonexistent in most woreda offices. In the near term, recording necessarily will be primarily manual, and records will be kept for some years on paper and film. Current conditions require investments in upgrading the quality of storage facilities. Due to cost considerations, the upgrading will have to focus first on woreda, rather than kebele, facilities. For the time being, the storage of documents and films will have to be concentrated in the former.73 Another major reason for poor data quality and lack of maintenance is the fact that the process for updating the records is difficult. In many cases, multiple visits to the records office are required. The time and travel required to do so are a real issue because, in most cases, records are updated at the woreda not kebele level. Based on the principles of providing services locally, transparently, with active community participation, with inbuilt safety mechanisms (backup), and with the registry book as the source of land information, suggestions have been developed to re-engineer the eight key registration processes. They are (a) first-level 72 Based on a recent detailed review of the registration system and development of recommendations through a stakeholder process (Orgut 2010e). 73 Orgut 2010f, provides a detailed review of the status of document storage in kebeles and woredas, and suggests a 5-year plan to upgrade storage. 79 certification, (b) second-level certification, (c) inheritance, (d) gift/donation, (e) exchange, (f) rent, (g) divorce, and (h) property formation. These re-engineered procedures should form the basis of improved service delivery. Possible strategies to improve the quality of existing records include: a. Undertake a large systematic program to validate and verify existing data and to update as necessary (the planned second-level certification program will do this but will take some years to complete). b. Undertake a smaller scale update process perhaps through the LACs (which were responsible for the initial compilation of the records) with substantial public consultation. c. Implement a program to update the records on a transaction basis (for example, the parcel areas could be updated by requesting that a survey plan be prepared before recording a request to update the records). d. Undertake a major PIA campaign to explain the processes and benefits of keeping the land records up to date. e. Re-engineer the process to make it simpler and more accessible to users. f. Use a combination of the above strategies. Some of these steps already are being considered, including the PIA campaigns and the re- engineering of the procedures. The strategy that might be adopted will depend on the quality of the records, which will vary from region to region and even woreda to woreda. There is benefit in undertaking a study of data quality so that a strategy and plan can be prepared to address the problems. An important point is that the implications of not having the data in the land records up to date will vary. Many land records systems have difficulty in recording succession. However, the implication of not recording succession can be limited to the increased risk of family/household disputes and the loss of revenue in transaction fees and taxes. Each of the four regions has a different design of a registry book and Book of Holdings Certificate (BoHC). A land administration system records information about three basic elements: (a) specified units of property (land parcel, apartment); (b) specified right holders (individuals, households, community groups, and legal entities such as companies); and (c) the specific rights held by the right holders over the property and the changes in these rights over time. The registry book in Amhara is well designed and suitable for registration activities. The registry book in Tigray records each holding on one page in the book but has room for information on only five parcels per holding on that page. The registry books in Oromiya and SNNP are holding based and have a major weakness in that they allow only one line in the book per holding. Both forms lack insufficient room to record all of the details of parcels in a holding. The book is virtually useless to record changes. The Ethiopian Strengthening Land Tenure Policy and Administration Program (ELTAP) has developed a parcel-based registry book. This book is simple to use and understand. The registry books in Tigray, Oromiya, and SNNPR should be redesigned, adopting either the holding-based structure used in Amhara or the parcel-based structure developed under ELTAP. In Amhara, the BoHC, which is prepared from the information in the registry book and given to landholders, is basically well designed. However, the BoHC could be improved by adding a unique serial number printed on all pages in the book and by adding a reference to the “holding type� in the book, particularly one that distinguishes between “private� and “communal� 80 holdings. The same improvements could be made in the other regions. A number of proposed changes that would benefit other regions also were identified.74 A parcel-based system is essential to chart holdings and establish and maintain the spatial framework to support the land administration system. Ethiopian rural land records pertain to holdings rather than to single parcels.75 This arrangement contrasts with most land administration systems in the world, which have adopted the land parcel as the specific unit of property. The land parcel can be mapped on cadastral maps and assigned a unique identification number. These numbered maps provide the spatial framework for the land administration system and enable a link to be maintained to right holders and documents lodged to effect changes in rights. Any change in the record system has the potential to create significant work in updating software and in re-sorting/indexing records and data. Much of the debate is due to the understandable reluctance to undertake this work. However, it is possible to introduce a parcel- based system in Ethiopia while maintaining the core elements of the holding-based system that has evolved in recent decades.76 A unique parcel identification system can be developed based on best practice that is unique, permanent, capable of change when the property is subdivided or consolidated, convenient for users and suitable for computerization, easy to understand and remember, and accurate. To maintain the holdings nature of the records, a holding index can be set up and maintained with a link to the parcel index. Urban land records are maintained by cities and municipalities under a great variety of systems. There are a number of initiatives to undertake large-scale orthophoto mapping projects in Addis Ababa and other major cities. Several cities are developing systems to record land information including information related to rights in land. There is a need to develop standards and guidelines for urban land records. Most of the procedural aspects of urban real property registration should be made available in the Urban Real Property Registration Proclamation. Norms and standards may be established implicitly by the proclamations. Where these need to be expressed explicitly, the responsible federal authority has the power to issue directives. 74 The specific proposed changes are for: Oromiya: (i) Increase the space for photographs so that a photograph of the spouse can be added, (b) increase the space to be able to clearly record changes such as the addition of a parcel to the holding, and (c) insert the holding number from the registry book. SNNPR: (a) Insert a holding ID, (b) insert a parcel ID, (c) add pages for updating changes such as the addition of a parcel to a holding, (d) add pages for second-level certification, (e) add space for holders to sign the certificates, and (f) add space for photographs in cases of more than one wife. Tigray: (a) The book should be for a holding, not simply a parcel (greatly reducing the number of books needed); (b) book should cover both first- and second-level certification; (c) full name of the spouse should be included; (d) a unique identification number should be included; (e) photographs of all holders should be included; and (f) sufficient space should be allotted for updates. 75 In Tigray and Amhara regions, the details of all the individual parcels in each holding are recorded together on the same page of the registry book. In Oromiya and SNNPR, only a summary of the holding is recorded in the registry book (because of the poor design of the registry book, allowing only one line per holding). In urban areas, documents are primarily permits and leases, and transaction records; and they refer to parcels and dwellings. 76 ELTAP developed an efficient parcel-based record system with a simple index linking parcel numbers to holding numbers. This linkage was essential because the registry Book introduced during the ELTAP Project has a page for each parcel, rather than a page for each holding. 81 Standards also can be ensured through the use of a customized registration application software system that enforces certain procedural, technical, and legal rules. Software is being, and will be, acquired and developed to operationalize the national urban real property registration and information system. Regions should adopt standard GIS software and then deployed it in municipal urban planning and land administration departments. Advice on software and hardware acquisition is provided in Annexes G, J, and K of DHV Consultants (2006b). Also recommended is phased implementation of computer systems (software and hardware), which would focus first on municipal registration; then, with the availability of more data, migrate to bigger and national systems. Specifically, no wide area network link is proposed initially, since real property data can be made available on disk to local, regional, and national users. Spatial data (maps) comprise three products: parcel maps (registry index map), urban topographic maps, and image maps.77 Another option is to delay acquisition of digital urban topographic mapping pending high-level policy decisions about the roles of Federal Urban Planning Institute (FUPI) and the Ethiopian Mapping Authority (EMA). Topographic mapping is not essential to real property registration. However, it is fundamental spatial framework data for integrating broad-based land information systems. Information, education, and communication are important aspects of a project that relates to property rights. Real property is crucial to individual livelihoods, the business environment, and socioeconomic development in general. Therefore, a public and stakeholder awareness and information campaign is an essential component of an effort to upgrade the urban land administration systems. 4.3.3 Systematic registration The rural certification program in the four main regions built on the records of prior land allocation and was undertaken in a very participatory process. Substantial certification work was completed in a very short time with a very low-cost approach (section 3.2). Policymakers set a timeframe for certification. The implementing agencies responded by adopting a two-level certification process: (a) first-level certification, which captured information on rights but provided very limited spatial information (the names of people with rights on adjacent land and an estimate of the parcel areas), and (b) second-level certification, which surveyed parcel boundaries and produced cadastral maps. Although substantial progress has been made in the four main regions, certification work has yet to commence in the emerging regions. In the latter, certification will begin with the development of guidelines or, in the case of Somali regions, proclamations. In areas in which first-level certification is not complete, once second-level certification is ready for scale-up, a decision may be made whether to follow the two-level approach or to implement a single certification process, the approach adopted in most other countries.78 A key resource constraint 77 The production and provision of these digital spatial data sets are detailed in appendix G of DHV Consultants 2006b. 78 If a single-stage certification is implemented, a strategy for prioritization and phasing will need to be adopted. If a decision is made to complete certification in a very short timeframe, the two stage-certification process is a clear response. In this case, the first-level certification process should learn from the experience in the four main regions and be modified to undertake certification in a manner that minimizes or eliminates the requirement for an expensive 82 to completing certification and updating the system is the limitation in staffing levels, particularly in Oromiya and SNNPR (Table 4.2). Communal land in rural areas generally has been registered in the name of the kebele rather than in the name of the group with rights over individual parcels of communal land. In the four main regions in Ethiopia, the registration of communal grazing lands has been undertaken as part of the ongoing first-level certification program. The boundaries of this land are not uniformly recorded in all regions, and the communal land generally is registered in the name of the kebele. However, the actual user rights for specific plots of communal land might belong to only group of households within a kebele or even from more than one kebele. Given this complexity and the fact that these groups are typically not a legal entity, no formal process to register individual plots of communal grazing land with specific local groups exists to date (Andersen and Dupuy 2009). In Amhara, the land proclamation provides for the devolution of rights in communal grazing land to user groups. However, guidelines for the consultative process and to support registration are not in place. Second-level certification of rural land has happened on pilot bases. The absence of a map implies a greater potential for boundary dispute. Thus, farmers have much interest in such upgraded issuance of certificates with maps (chapter 3).79 The USAID-funded ELTAP program (2005–08) included mapping and registration in 24 kebeles in the 4 main regions. Although extensive surveying and registration had taken place, certificates eventually were issued to only a relatively small number of parcels within the pilot areas. The methodology used in this pilot relied on hand-held GPS receivers (costing approximately US$100 each), and computerized storage of data and maps. In assessing the lessons from the pilot, a USAID evaluation team pointed out that the hand-held GPS methodology “is adequate to identify and describe a plot of land in terms of its shape and size, but it is not accurate to define a boundary to the level of accuracy required by farmers� (USAID 2008). The results point to the need of setting certain standards of accuracy and verifiability of second-level certificates. A subsequent assessment of the ELTAP experience highlighted the ability to achieve results at a relatively low cost, and the capacity of this approach to produce cadastral index maps (CIMs), which are essential for second-phase certificates (Orgut 2010b, 16). However, ELTAP has not produced cadastral index maps and the way the individual parcels have been surveyed under the ELTAP pilots would create difficulties in forming cadastral index maps. Concerns about inaccuracies induced the Amhara region to reject the hand-held GPS methodology. The other three regions accepted the pilot, but very few certificates have been issued under the project. Visits to Tigray, Oromiya and SNNPR in March 2011 and consultation with government staff has confirmed that the procedures developed under ELTAP are not scalable. Nevertheless, further attempts are made under the follow-up ELAP to improve procedures in a smaller geographic area and the government of SNNPR has provided ETB3 million during the current fiscal year to adopt this approach in 13 additional woredas. As noted in the Concept Note for LALUDEP, additional work is required to reach consensus on the appropriate approach and methodology for second-level certification. second-level certification activity. Options here include the use of provisional certificates that might require the landholder to bear the cost of upgrading the certificates at some time in the future. 79 Indicative willingness to pay is significant, but nevertheless amounts to perhaps only 10% of the likely cost of second-level certification. In other words, the costs of certification would have to be substantially borne by the government. 83 The challenge has been to agree on an appropriate, cost-effective methodology that can significantly improve the first-level certificates. There is still no full consensus on the common definition and best approach to second-level certification. Concerns regarding the advisability of taking up a second-level certification program on a full national scale are also valid. The causes for doubts include the inaccuracies in the rural land registries, generally inadequate status of urban cadastral maps, incompleteness of first-phase process, limited availability of skilled staff, and high cost of implementing a second-phase certification program, even under optimistic scenarios.80 These considerations need to be weighed compared to the presumed benefits of upgraded certificates, namely, additional reduction in land disputes, further enhancement of perceptions of tenure security, greater efficiency of the land market, and greater fairness in tax assessments. Reaching consensus on how to move forward with second-level certification is likely to require a review of recent Ethiopian and international experience, some stakeholder consultation, and possibly experience from additional pilot activity utilizing different technologies and methodologies. A brief note to guide the thinking is presented in attachment 7. The likely approach for a second-level certification process would be to rely on a broad low-cost solution such as the use of orthophoto maps, which will be cost effective for most rural land parcels. For areas not identifiable on orthophotos, GPS systems can provide submeter accuracy.81 More accurate and more costly ground survey techniques such as real-time kinematic (RTK)/GPS and total stations also can be used in cases for which the higher accuracy and costs can be justified, as in urban and periurban areas. The major costs of these approaches are the orthophoto maps. However, these orthophoto maps have broader applications and wider benefits than land administration alone. Consequently, most countries, including Romania, Rwanda, and Thailand, have relied on large-scale orthophoto maps for major systematic registration programs.82 Box 4.1 Cadastral Index Mapping Pilot and Computerization in Amhara The Amhara National Regional State, with support from the World Bank and UN Agency For Human Settlements Providing Adequate Shelter For All-Global Land Tool Network (UNHABITAT-GLTN), has piloted the preparation of cadastral index maps (CIMs) using a high-resolution satellite image (HRSI) in Zembella kebele in Fagita Lekoma woreda. The process comprised screen digitizing, field checking, verifying data through kebele LACs, establishing field control points, developing a software for local geo-referencing (named “kebele calc�), transforming data using the kebele calc, and comparing field and registered data. Later, the results were linked through the locally developed Information System for Land Administration (ISLA) software. The test was based on an ortho-rectified 80 In fact, the uncertain benefit-cost advantages of second-phase certification led Orgut (2010a) to question the viability of such a comprehensive program in the near term and to suggest instead a more gradual approach. In it, inaccuracies and deficiencies in the existing land registries would be tackled first; technical, managerial, and administrative capacities are enhanced; legal gaps are corrected; and limited second-level certification efforts are focused on high-priority areas (Orgut 2010a, 28; Orgut 2010b, 8–9). The latter state: “A fundamental question needs to be answered, ‘Are the benefits worth the cost?’ In view of some of the problems which exist with first-level certification... it may be better to spend less money on improving existing paper based systems and data, rather than creating more (and more complex) data for the 2nd level.� 81 Omnistar or GPS with EGNOS correction are two systems that provide submeter accuracy. 82 A system using hand-held GPS units has been developed under ELTAP and demonstrated a positional accuracy of approximately 7–10 meters. This technology is low cost and has potential application in Ethiopia. However, the low accuracy limits the application for small parcels such as those in urban and periurban areas, and produces data that has limited benefits in re-establishing boundary points. 84 QuickBird Image. While the original positional accuracy of the ortho-rectified image was nearly 12 meters, the accuracy was increased to below 1 meter through locally established control points and the support of the kebele “calc� software. The costs of the process are US$3–$5 per parcel, including the preliminary establishment of surveying infrastructures and the cost of the image (proportional for Zembella, since the HRSI covered the entire woreda). The pilot process was built on the first-level certification of the region. Hence, it was participatory. These participatory aspects of the exercise are expected to deepen with the involvement of individual right holders in assessing CIM results and sharing their views and complaints during public hearings. Once refined, this “second-level� process piloted in Amhara may well take the certification and registration activity to a next level. Results based on only one kebele obviously must be interpreted cautiously. Nevertheless, the pilot has shown that the production of HRSI-based CIMs should be considered as one of the options for developing a cost-effective means of generating spatial framework. Whatever the approach chosen, second-level certification requires prioritization and phasing and will take time. A 15-year horizon would enable a gradual approach. The pace of second-phase certification would pick up in later years as legal issues are ironed out; effective organizational structures are put in place; capacities are enhanced; and the ICT infrastructure is upgraded. In this view, “…urban areas should be the first priority for land administration services because [they are the places in which] the returns on the investment will be highest and [in which] the poverty reduction effects [are likely] to be most significant… .� (Orgut 2010a, 28) The same logic also implies that priority should be accorded to upgrade certification in the rural areas whose land is of high value due to favorable agricultural potential, or to land in periurban areas, and in proximity to major roads. The total cost of implementing a second-phase certification program is significant and varies tremendously by the approach chosen. The costs (including base mapping, survey, adjudication, registration, and certification of nearly 50 million parcels) have been estimated at US$174 million–US$490 million, or US$3.52–$9.88, respectively, per parcel (Table 4.6). Actual costs between these two figures will depend on implementing the survey methodology and the extent to which first-level certification simplifies the second-level adjudication and registration.83 A project of this size will take time to design, plan, and organize, particularly to build up the required human resource capacity. A 10– or, more likely, 15–year program will be needed to do these activities properly.84 83 Much uncertainty prevails regarding the accuracy of the land records due to the incomplete updating of the registry in many localities. All of the 4 scenarios above assume management and capacity building costs of US$20 million. This figure is based on a 5-year capacity building program managed by international contractors with local counterparts costed over 10 years. 84 In Thailand, a country of some 65 million inhabitants, significantly more arable land than Ethiopia, and a much higher per capita GNP, it took over 20 years to complete a nationwide land-titling program that issued certificates with spatial information and mapping. Given technological advances and the experience of rapid first-level certification in Ethiopia, faster progress can be expected in Ethiopia. 85 Table 4.6 Alternative Cost Estimates for Second-Phase Certification Cost per parcel (US$) Orthophotos (full Registra- Total Scen- Survey Description coverage in all tion and cost ario and adjud- Total scenarios) certifica- (US$m) ication tion High cost: Full-phase 2- 1 1:5,000 scale 6.0 3.0 9.9 491 level certification Mid-cost: Full phase 2- 10% 1:5,000 scale 2 4.0 2.0 6.6 325 level certification 90% 1:10,000 scale Low cost: Full phase 2- 3 1:10,000 scale 2.0 1.0 3.5 175 level certification Sporadic registration: Costs based on 10% of 10% 1:5,000 scale 4 parcels registered; 6.0 2.0 13.5 67 90% 1:10,000 scale periurban areas as priorities Source: Orgut 2010b. The higher estimate of approximately US$10 per parcel for second-level certification compares well with international experience (Table 4.7) including with the more recent experience in Rwanda. In a comprehensive pilot program, Rwanda developed a field process to demarcate and adjudicate rights in land, using orthophoto maps as a map base, and to issue leases to landholders at unit cost of US$9–$11 per parcel. With support from DfID, the government is scaling up this process and plans to complete the systematic registration of the entire country (8– 11 million land parcels) in a relatively short period. The government has nominated a period of 3 years (2009/2010–2011/2012). Achieving the target in terms of demarcation and adjudication is feasible, but doing so in terms of lease issuance/registration is unlikely, due to constraints in institutional capacity and in timely and adequate resource allocation. The scale-up commenced only in the past year. However, early indications are that leases can be produced systematically at a unit cost of approximately US$7. Table 4.7 Unit Costs of Systematic Registration in Selected Countries Armen- Kyr- Moldo- Indone- Thai- El Peru Peru ia gyzstan va sia land Salvador (urban) (rural) Pre-Field 4.89 Geodetic network - - 5.66 - 0.39 Cartography 0.20 - 7.08 7.05 0.24 11.26 Compilation of existing records 0.02 0.03 1.53 1.30 Publicity campaign 0.02 0.31 0.55 1.94 0.42 Acquisition of gov. equipment 0.68 0.91 - 1.50 Field 19.32 Collection of claimant 1.00 0.30 3.77 0.23 3.62 information Boundary investigation, survey, 4.57 2.09 7.64 9.67 1.61 10.50 marking Conflict Mitigation - - - 0.06 0.08 Post-Field Quality control 0.12 0.14 0.94 0.058 10.00 Legal validation 1.00 0.15 0.56 Public display of field results 0.02 - Conflict resolution - - 0.02 Prepare land record 1.00 0.04 2.92 2.89 86 Armen- Kyr- Moldo- Indone- Thai- El Peru Peru ia gyzstan va sia land Salvador (urban) (rural) Prepare cadastral maps/plans 0.82 0.04 1.98 1.44 2.37 1.68 Cadastral/registry database 0.50 1.06 3.77 design Data entry 0.10 0.03 0.19 Register property rights in 0.05 0.14 7.55 5.44 registry Issuance of titles to beneficiaries - 0.01 0.94 1.95 Administration/management 3.25 5.30 1.89 3.89 7.27 9.28 Total costs per parcel 13.25 10.55 46.41 16.30 24.21 29.74 12.68 55.69 Amount paid by beneficiary - - - - 2.55 - - - Total costs 13.35 10.55 46.41 16.30 21.66 29.74 12.68 55.69 Source: Burns 2007. Registration of rights that includes a spatial reference is even more urgent in urban areas than in rural land. While starting with sporadic registration, Ethiopia needs systematic adjudication and registration of urban properties, which will entail regularizing informal settlements and properties. Several cities in Ethiopia have attempted to compile cadastres for various purposes. Most of these efforts have failed to produce sustainable results. Reasons include a focus on the technology rather than a clear definition of the requirements and necessary work procedures; a lack of engagement by city officials in specifying and “owning� the installations; and limited capacity to maintain and operate the systems.85 The MoUDC is developing a road map for land administration in urban areas and for major cities a central strategy is to build on the experience on Addis Ababa, which is presently implementing a major land information/cadastre project. These efforts and experiences will information the future of land administration in the urban sector but the needs are great and steps can be taken to improve systems. Regularization requires advance contemplation and preparatory legal foundations (to determine criteria for eligibility and various aspects of the procedure). The three highest priorities are: a. Complete the urban parcel maps by more expeditious means (for example, by using large-scale orthophoto maps produced from aerial photography rather than by accurate but costly ground surveys) b. Where appropriate, using a unique parcel identification code to index all property, tax, and planning-related information c. Index all property records according to parcels, not persons. With the advance of registration, cadastral maps can be updated and improved with the addition of other land information aspects, such as encumbrances and mortgages, and land use limitations.86 Advance thought needs to be given to procedures for recording changes related to transactions and life events. This effort will require extensive strengthening of staff numbers and skills at the municipal level.87 85 The recent effort by Amhara to complete a cadastre for Bahir Dar and for Gondar is an example. The system was designed by the regional authorities with little involvement by the city in the specifications and very limited technical capacity. Much of the work to develop the system has been completed but the cities do not have the capacity to implement. 86 Detailed advice on urban surveys is provided in appendix G, DHV Consultants 2006b. 87 Detailed advice on staffing requirements and training needs is provided DHV Consultants, 2006b. 87 4.3.4 Information strategy and implementation Despite the initial focus on improving the manual land records systems, there are benefits in starting to develop an overarching ICT strategy as it will: a. Standardize procedures and strengthen quality control b. Protect against fraud by creating audit trail of changes of entry c. Protect interests when kebeles are moved from rural to urban regions d. Achieve economies of scale by adopting standard procedures, platforms, and software products e. Clarify how different technologies may be used to create records and record information about how these records have been produced f. Create the GIS cadastral fabric, which will produce valuable spatial resources for use by planners, utilities, and local authorities g. Facilitate the secondary market through deriving additional products and services, such as land transactions/leasing activity, market transfer rates, and identification of areas for investment (Orgut 2010c). The proposal for how ICT could support the technical unification of the rural/urban sectors is shown in Figure 4.1. This proposal provides for a 6-fold development phasing, giving priority to the first two components: the “land rights� management system and the parcel management system. Using a simple set of assumptions, a highly provisional estimate of the first 2 phases, including the development of software, suggests that the implementation will cost some US$17.8 million. Of this amount, an estimated US$11.8 million would be needed to cover equipment/systems and the balance to cover technical assistance (national and international). Figure 4.1 Proposed ICT Strategy Phase Rural Urban Town/ Small Subsystem Region Zone Woreda Kebele City sSubcity town Input “Land Rights� 1 management Output system Query Input Input Parcel 2 management Output Output system Query Query Technical 3 building register Valuation 4 subsystem 88 Phase Rural Urban Land use and environmental 5 protection subsystem Advanced 6 products and services Source: Orgut 2010c. Along with the development of an ICT strategy, and the gradual increased reliance on digitized data and computerized record and data storage and management, the software to facilitate various aspects of land administration needs to be developed and/or adapted. There are a number of alternatives that might be considered by the regional governments, including: a. Directly commission further development of ISLA software b. Commission a commercial system integrator to design and develop new software based on these specifications suitable for use in all regions c. Use a capacity building approach in which an external consultant/system integrator works directly within a new regional ICT unit to design and develop new software, or extend the existing ISLA software d. Tender a single software development procurement package including design and development, installation, training, and data conversion for installation and adoption across the organizational structure within a region. Of these possible approaches, options 1 or 3 are recommended because they involve region staff directly in the development process, thereby enhancing knowledge transfer and skill development (and promoting sustainability). Furthermore, these approaches maximize the use of the existing ISLA software so are likely to cost less. Certain steps in the development can be done centrally. However, ultimately, the adaptation of software needs to be led by regional governments because regions differ in population density and tenure patterns, and therefore have different needs. Urban and rural aspects also may require specific adaptations due to the differences in business processes. The development of the software and the cost of developing the software have been included in the ICT implementation plan and budget set out above. 89 5. THE WAY FORWARD An effective land administration system is critical for Ethiopia’s development. Sustaining Ethiopia’s path of rapid economic growth and progress toward achieving the Millennium Development Goals and the targets of the GTP will require, among other things, continuously strengthening the enabling environment to continue the country’s transformation from predominantly subsistence agriculture to commercial farming, industrialization, expanded services, urbanization, and further improvement of the rights of women and vulnerable people as well as good governance. A major element of this enabling environment is a fair, transparent, and effective land administration system to provide tenure security and facilitate the efficient and sustainable allocation of Ethiopia’s land resources and its labor and capital. Strengthening Ethiopia’s land administration system must be based on a solid understanding of the historical, legal, and institutional context and current practices. This report has tried to set this context in chapter 2. Also required is an understanding of the significant achievements made with respect to land administration and remaining challenges (chapter 3). Based on that discussion, chapter 4 discussed the various components of land registration and some other elements of land administration and provided options to strengthen it. The report concludes that Ethiopia has significant scope to further develop the legal system; to strengthen implementation capacity and the organizational set-up; and to improve actual land registration services. Areas for legislative strengthening are related, particularly, to urban land, including the unification of the currently dual urban system of urban land rights (“permit� and lease), standards for a minimal land plot for residential development, provision of strong legal basis for certification and registration; expropriation and compensation, not the least with an objective to reduce the need for expropriation and replace it through techniques based on voluntary participation of land tenants (“land re-adjustment�, “land pooling�); and further fine-tuning rural land legislation and adoption of legislation in all regions. Ethiopia also should consider its organizational structure for land registration, particularly the coordination mechanism for rural and urban land; further strengthening the human and physical capacity for land administration and land-use planning at all levels; and developing clear and transparent ownership records in urban areas including through the introduction of more sustainable policies and practices related to allocation of new plots for housing and better pricing of urban land. Areas of improving land administration services include establishing a complete land registry system and maintaining it; and gradually rolling out a second-level certification program; while working on harmonizing and coordinating the services across the country and across urban and rural areas. Moving forward, assignment of clear responsibilities and prioritization are crucial. While this report identifies the above areas for further strengthening and provides options of how to address them, one must take account of which stakeholders are involved in such actions and realize that not all issues are equally urgent or can be dealt with immediately and effectively. To prioritize them and provide a sense of urgency, the action matrix presented at the beginning of this report clarifies responsibility for various areas of strengthening. The matrix provides detailed recommendations for how to address the existing weaknesses. Given Ethiopia’s federal system, the responsibility for the implementation––as well as the financing and development of an eventually self-financing system––of a land administration system rests largely with the regions, municipalities and woredas. The capacity of these levels should be further enhanced and the 90 autonomy, especially of municipalities, further strengthened. However, the federal government (and the donors that are asked to and want to support it) can play a very significant role. The proposed LALUDEP, for which a concept note already exists, seems the adequate tool to implement the reformed and strengthened land administration system. The concept note already emphasizes the strengthening of the capacity of the federal government itself to enable it to develop adequate policies, legislation, regulations, and manuals; and to be able to provide technical assistance and monitor the performance of the regions (and woredas). The support to regions and woredas may be provided on an on-demand and performance basis. 91 Attachment 1. Organization of Land Administration in Selected Developing and Transitional Countries Russian Federation When the land reform activities started in the Russian Federation following the break-up of the Soviet Union after 1991, there were different organizations for rural and urban land and a separate mapping authority.88 Agricultural land came under the Department for Land Administration within the Ministry of Agriculture. Practically all urban and settlement cadastral activities were performed by Buro Tehnicheskoi Inventarizacii (BTI). Municipal organizations were responsible for house and apartment registration in the towns. In 1990 the Russian Federation State Committee for Land Reform was created as a separate governmental body to register land. At that time, the regions and towns developed their own systems for land registration. The duplication was partly the result of weak federal organizations and the lack of coordination of donors, combined with the regions’ independent status. Because the duplication was costly, the government requested harmonization of the system. In 1992 the State Committee for Land Reform (Roszemcadastre) was designated as the main actor for rural land at the federal, regional and municipal levels. The committee was made responsible for the development of the extensive land reform in Russia and the creation of a unified national property cadastre for rural land. The primary objectives were to strengthen the federal organization to take the lead in this development and to establish reliable registers that included property information and cadastral maps. Over time, it became apparent that the activities at the federal level also must include legal development, land valuation, organizing training, and skills transfer. At that time, plots were surveyed primarily by the authorities and partially private companies that worked in the towns. In 2004 to establish a common infrastructure for Land Information, Roszemcadastre (Rural) and Buro Tehnicheskoi Inventarizacii (Urban) were merged into the State Agency, Rosnedvizhimost, under the Ministry of Economics. Other stakeholders––such as the Ministry of Justice, responsible for the registration of rights to properties, and the Ministry of Taxation and Fees, in charge of property taxation– –were actively involved in the development of the Land Information System. Other core actors were the municipalities, responsible for land-use planning, municipal engineering, and environmental protection; and credit institutes, which granted loans against real estate offered as collateral. The result was the establishment of a land information infrastructure, which consisted of core data of common interest to the three main actors. At this time, surveying and mapping were under a separate governmental body. In March 1, 2008, a new cadastral law came into force that stipulated that the cadastre was to be operated throughout the entire Russian Federation using uniform technology. The new law would merge the separated building and land cadastres and the merging of the land register and the cadastral authorities. In 2009 the Federal Service for Government Registration, Cadastre and Cartography was created under the Federal Ministry of Economic Development. The new organization, which included the Federal Agency of Geodesy and Cartography, is responsible for three processes: 1. Setting up a unified system of cadastral records for immovable property 2. Registering rights to immovable property and related transactions 3. Organizing the spatial data infrastructure of the Russian Federation and managing government real estate land and its privatization. As evident from the review above, a registration system for rural and urban land and rights to land under one authority was under discussion almost from the start of 1990, but it took almost 20 years to merge the rural and urban land registration systems into one organization. 88 Material in this appendix is adapted from Orgut 2010a. 92 Vietnam In 1993 the National Assembly approved a Land Law, which established the General Department of Land Administration by merging and reorganizing two state bodies: the General Department of Land Management and the National Department of Surveying and Mapping. The land administration system is now a nationwide four-tier structure with (a) the General Department of Land Administration at the central level; (b) the Department of Land Administration at the provincial (state) level; and (c) the Bureau of Land Administration at the district level belonging to the People’s Committee with (d) 1 or 2 Land Officers in each commune. Urban and rural lands are handled by the same administrative unit. The land administration bodies are responsible for land administration and survey-mapping as well as the following 8 tasks: • Preparation of land legislation and policies for submission to the authorities • Cadastral mapping, including the preparation and update of the land register and issuance of land tenure certificates • Land valuation • Land dispute resolution • Establishment and maintenance of the Geodetic Reference Network • Management of remote sensing materials (aerial photos and satellite imagery) • Preparation of topographic base maps • National land database management. South Africa South Africa has a federal system of government. Deeds registration, land survey, and land reform are the responsibility of the national (federal) government. The Surveyor-Generals’ offices examine and approve all cadastral surveys, and compile and maintain plans showing the relationship of the various parcels of land to one another. The Registrars of Deeds register land rights and responsibilities and maintain records relating to all registered parcels of land, rural and urban. The control of these operations is vested by statute in the (federal) Minister responsible for Land Reform, who ensures uniform country-wide standards of surveys and deeds. Provincial governments have responsibility for a number of functional areas closely related to land reform. They are mainly the areas in which national and provincial governments have concurrent responsibilities. These areas comprise agriculture, environment, soil conservation, housing, regional planning, and urban and rural development. Local governments also have constitutional functions that affect land use and planning. In addition, traditional authorities also carry out land-related functions using customary law. 93 Attachment 2. Rural Land Administration System in Different Regions of Ethiopia Amhara Land administration matters are handled by the Bureau of Environmental Protection, Land Administration and Use (BEPLAU), reporting directly to the regional government. The bureau is responsible, among other things, for carrying out land registration and certification, conducting cadastral surveys, establishing a data base for regional land administration, land use, and environmental activities, and property valuation for assessment of compensation. Staffing levels for land administration in Amhara are the most complete among Ethiopia’s regions (27 assigned at the central level, of 30 required; and 93 assigned at the zonal level, of 124 required; additional recruitment ongoing). The situation is different at the woreda level, in which approximately only 50 percent of the 1,536 required positions are staffed. Amhara is the only region that has put in place staff responsible for land administration, filling a reported total of 920 of 3,146 positions. Various land administration training programs are ongoing, but skills at the woreda level are limited, especially for second-level certification activities. Oromiya The Oromiya Bureau of Land and Environmental Protection (BoLEP) was established by Proclamation 147/2009, which aimed to create a single executive organ to administer and regulate urban and rural land, and to prepare land-use plans for the region. The head of BoLEP reports to the head of the regional government. Within the bureau are four process leaders: Rural Land Administration, Land Use, Urban Land Administration, and Environmental Protection. Presently, in rural woredas and kebeles (administrative villages), land matters are handled by the local government apparatus. Some officials handle specifically land issues (such as registration and approval of rentals); while elected organs (LACs) deal with preliminary adjudication processes, some aspects of the land certification program, and some dispute resolution matters. Bureau personnel are working to merge the urban and rural land administrations to create a unified land system. A regional proclamation on this matter is under preparation. Nonetheless, the merger is likely to entail difficulties. BoLEP officials reportedly have reported problems, including land-use conflicts and uncontrolled expansion of urban development on rural land (particularly overspill from Addis Ababa) (Orgut 2010a, 12). Given BoLEP’s limited capacity to handle urban land management issues and the lack of base maps for the cities, urban land administration is perceived as the bureau’s biggest challenge. In an attempt to develop a cadastral system, various cities individually have hired consultants. This effort has proven ineffective, because the cities have insufficient capacity to maintain the systems that have been developed. Capacity problems, scarcity of trained staff, and limited transportation resources also afflict BoLEP. Within Oromiya 363 towns and 7 urban structures generate sufficient revenue, for the most part, to finance their own land administration. However, the region also has 18 zones, 264 rural woredas, and 6,419 kebeles, which require a significant budget for land administration in the absence of adequate revenue generating activities. On average, each zone has only 10 land administration specialists. Southern Nations, Nationalities and People’s Region (SNNPR) Rural land administration is the responsibility of a Land Administration and Use Team, under the head of the Bureau of Agriculture and Rural Development. The team also includes an environmental expert who handles both urban and rural areas. A draft regional proclamation, which will establish a separate Bureau for Land Administration and Use reporting directly to the regional government, is expected to be 94 approved soon. Environmental protection is expected to be set up as a separate authority as well. The region has 13 zones, 8 special woredas, 126 woredas, and 4000 kebeles. This region is home to 56 different nationalities each with a distinct culture and language, thus further complicating land administration and certification activities. Moreover, staff levels and land administration skills at the zonal levels are grossly inadequate. Even the existing 8 staff members of the land team at the central level do not at present have any land administration skills. A recent assessment of land administration capacities in Ethiopia concluded that SNNPR lags far behind the other three regions, due primarily to extremely inadequate human and material capacities (Orgut 2010a, 26). Tigray Rural land administration in Tigray is under the Environmental Protection, Land Administration and Use Agency (EPLAUA). Nominally, the agency is accountable to the regional Bureau of Agricultural and Rural Development (BoARD). However, in practice the interaction is one of liaison, as the agency has an independent budget from the regional government. At the woreda level, EPLAUA staff report to the woreda administrator (previously they reported to the head of the ARD office). While a proposal to integrate urban and rural land administration in Tigray has been floated by EPLAUA officials, the plan has not yet been approved. At present, the agency has three core processes: Land Administration and Use, Environmental Protection, and Support. While Land Administration and Use should have approximately 28 experts at the regional level, only 50 percent of this number are assigned (of whom 7deal with land administration). In Tigray’s 34 rural woredas, there are on average 5 EPLAUA staff per woreda, compared to the target number of 9. As in Oromiya, elected LACs at the kebele and subkebele level perform some land administration functions, primarily dispute resolution. Material and equipment at the central level are insufficient (only one vehicle is available to facilitate support and operational monitoring). 95 Attachment 3. Assessment of First-Level Rural Land Certification in Ethiopia Attachment 3 Table 2 Wealth and Gender Characteristics of the Rural Land Registration Program Topic Total Rich Poor Amhara Oromiya SNNPR Tigray Wealth aspects Encountered encroachment on public land (%) 38 N/Aa N/Aa 56 28 22 88 Family member attended pre-program informational 81 84 80 87 79 85 72 meeting (%) Consider selves well informed (%) 78 79 77 82 73 82 80 Household member in LAC 11 12 10 9 12 14 7 Had some land document 59 56 61 84 39 54 93 • Preliminary Registration Paper 39 46 44 55 40 15 31 • Book Of Holding/Certificate 60 54 56 44 59 84 69 Parcel not yet registered 10 12 6** 8 11 14 6 Plot measured by eyeballing/pacing 24 20 25** 32 25 5 17 Gender aspects Female attended pre-program meetings 38 35 43* 38 32 44 51 Female participated in LAC election 17 15 20* 21 13 20 23 Kebele has female member on LAC 20 N/A N/A 33 8 30 25 Female knows where certificate is kept 86 85 89 90 81 85 85 Certificate in husband’s name only 36 34 36 9 58 21 71 Both husband’s and wife’s names on certificate 52 49 49 79 29 69 13 Source: Derived from Deininger and others 2008, tables 3 and 4. Note: a Richest and poorest quintile, respectively; asterisks indicate the significance of differences between the two groups: ** = significant at 1%; * = significant at 5% ; N/A––not available. 96 Attachment 3 Table 2: Willingness to Pay (WTP) for Land Certificates Topic Total Richa Poora Amhara Oromiya SNNPR Tigray WTP to replace lost certificate (ETB) 12 20 8 9 22 7 5 Will pay for a certificate (if do not have yet) 93 97 86** 88 94 93 92 Would like to add map/sketch (%) 90 90 90 88 95 86 93 Willing to pay for adding a map or sketch (%) 93 95 88** 89 95 95 92 How much WTP second-level (ETB) 13 N/A N/A 10 16 12 5 Source: Derived from Deininger and others 2008, tables 3 and 4. Note: a Richest and poorest quintile, respectively; asterisks indicate the significance of differences between the two groups: ** = significant at 1%; * = significant at 5% ; N/A––not available. 97 Attachment 4. International Experience in Land Administration Phasing Land administration projects typically are long-term projects implemented over many years. Land administration projects such as those in Indonesia and Thailand were planned as long-term projects implemented in a number of 5-year phases. Indonesia The situation in 1993 in Indonesia provided a weaker foundation for a program to strengthen land administration. Following 12 years of preparation, the Basic Agrarian Law was introduced in 1960. However, by 1993, only 20 percent of nonforest land was registered. The media regularly highlighted problems such as corruption, multiple certificates for the same parcel, public mistrust in the land administration system, and conflicts between formal and traditional land administration practices. Sporadic registration in the formal system was not servicing even the predicted demand due to increasing population. To address this situation, a 25-year program was prepared, to be implemented in 5 phases of 5 years each. Based on early tax-mapping records, it was estimated that, at the end of the 25-year period, the total number of parcels in Indonesia would be 78 million. The nature of the planned phasing is set out in Attachment 6 table 2. Implementation has not gone as planned; the first phase was extended to 7 years. However, the output for the first phase of US$1.957 million has exceeded the planned target of $1.2 million. Due to a range of factors, there also was a delay in implementing the second phase. Attachment 6 Table 2. Planned Phasing of Land Administration Strengthening Activity in Indonesia Planned Phase Period Scope output Phase very much an institution-building phase. Significant work on policy framework. Systematic registration activity confined to Java. Project areas 1994– 1 1.2 mil selected on basis of assisting in development of efficient land markets and 99 alleviation of social conflict over land. Project focused on offices receptive to change and keeping geographic spread of activities manageable. Phase will build on processes and procedures developed in first phase. Major part of systematic registration output would still concentrate on Java, area of most demand, but activities would be carried out to test and refine procedures 2000– 2 6.0 mil to register communal adat (for example, in West Sumatra). If socially 04 acceptable, pilots could be conducted in South Sulawesi. Additional work would be required to strengthen the National Land Affairs Agency of Indonesia as an institution with automation, computerization, HRD, and training. Phase would concentrate on islands of Java and Sumatra. Work could 2005– commence in South Kalimantan on basis that efficient procedures have been 3 11.0 mil 09 developed to mark forest boundaries, reclassify land, and incorporate customary tenure procedures. Work in this phase also would concentrate on Java and Sumatra, with 2010– 4 13.0 mil increasing activity in outer islands based on results of social assessment and 14 clear selection criteria. 2015– Phase would complete planned 25-year program. Activities would be 5 13.0 mil 19 undertaken in most remote provinces, subject to social assessment. Source: BPN 1993. Thailand Four phases were implemented in Thailand. The Thai project built upon a strong legal and policy framework with the initial emphasis on increasing capability to undertake systematic registration and the geographic expansion of systematic registration activity. In later phases, improved service delivery was emphasized. This change in emphasis can be seen in Attachment 6 table 1. There also was a geographic spread in systematic titling activity, with the initial phase concentrating in the lower North-East of 98 Thailand, the poorest provinces in the country at the time, and in the North of Thailand, an area with potential for economic growth. The second phase continued the mix of economic and social objectives, with extensive work in the Central and North East as well as in the Eastern Seaboard, an area targeted for economic development. The third phase completed the work in the North, North-East, and Central regions. The fourth phase filled the gaps and concentrated in the South. Attachment 6 Table 1. Thailand Land-Titling Project Component Structure Actual cost Base cost Item Component-Phase I (output 1,634,533 titles) (US$mil) (%) 1 Rural mapping, surveying, and systematic adjudication 37.8 60.9 2 Urban mapping 2.8 4.5 3 Land administration (including civil works) 6.0 9.7 4 Valuation 0.7 1.1 5 General institution building (including technical assistance) 14.8 23.8 Total, Phase I 62.1 Actual cost Base cost Item Component-Phase II (output 2,100,377 titles) (US$mil) (%) 1 Cadastral mapping and remapping 25.6 29.9 2 Land titling and administration 49.9 58.0 3 Valuation 0.6 0.7 4 Institution building 4.2 4.7 5 TA and training 5.5 6.4 Total, Phase II 85.5 Base cost Base cost Item Component-Phase III (output 4,772,055 titles) (US$mil) (%) 1 Land titling (including surveying, mapping, and title issue) 118.9 67.8 2 Improved service delivery 17.1 9.7 3 Strengthening DOL 17.5 10.0 4 Valuation 15.1 8.6 5 TA and training 6.3 3.6 6 Studies (socioeconomic and environmental impacts) 0.5 0.3 Total, Phase III 175.4 Source: Rattanabirabongse and others 1998. 99 Attachment 5. Developing a Spatial Framework for Land Administration A. Spatial Framework for Land Records Efficient systems to record rights in land need to manage: • registers of largely textual or alphanumeric data that record rights in land and/or transactions or changes over time in rights in land; and • a spatial framework that define the extent of land over which these rights apply. In many countries there is a weak or non-existent spatial framework in the systems that records rights in land. This typically results in problems such as overlapping claims and disputes over boundaries and can be a major cause of uncertainty in rights in land. There are a number of alternative approaches to establishing a spatial framework for land records. These approaches include using local, generally understood references such as street addresses, using abuttals to adjoining property or undertaking surveys connected to local or national coordinate systems. In practice more than one approach may be used as a spatial framework for land records in a jurisdiction and typically frameworks evolve from simple approaches such as local isolated survey plans, local reference systems, or abuttals to more comprehensive frameworks such as comprehensive surveys connected to the national coordinate systems (a “cadastre�). There is no universal definition of the “cadastre� and many different approaches. In Australia the cadastre has been compiled by coordinating accurate but isolated surveys. This coordination was initially undertaken to graphical accuracy and is being upgraded to survey accuracy. In Demark the cadastre is based on isolated islands of maps covering a village and surrounding cultivated areas that have been incorporated into a national system based on the national coordinate system. In England and Wales the spatial framework for the title register is large scale maps derived from the national topographic map base prepared by Ordinance Surveys. In Thailand large scale maps complied using plane survey methodology provided the initial spatial framework for the title register. The title register expanded as the spatial framework expanded. New technology such as photogrammetric maps and satellite positioning strengthened and expanded the geographic cover of the spatial framework. The frameworks have usually started by addressing specific requirements in specific localities. Where a decision has been made to expand the geographic cover of the system, this was generally undertaken in a manner that sought to expand the cover as extensively and as quickly as possible even if this meant compromises in the accuracy of the spatial information. Programs to improve the accuracy of spatial frameworks have generally been improved over time in response to clear user needs rather than as major jurisdiction-wide programs. B. Requirements for a Spatial Framework There are two broad aspects to the spatial framework that support the recording of rights in land. The first is a topological or indexing aspect that supports a range of applications, including: • the validation and verification of registered/recorded land, including the identification of duplicate or missing records and the identification of possible problems with overlapping parcels • a spatial framework for data queries and access to the data in the register; • identification of parties with an interest in a particular land parcel for a range of purposes including the identification of adjoining owners for service of notice; and • the identification of land parcels recorded in the register, including support for the sub-division or consolidation of land parcels. The second is a metric or calculation aspect that defines boundary corners and supports a second set of applications, including: 100 • strong evidence to support the resolution of disputes over boundaries; • the accurate re-instatement of parcel boundaries; • the calculation of accurate parcel areas, offsets, etc.; and • the accurate determination of updated parcel dimensions where land parcels are sub-divided or consolidated. Most jurisdictions have concentrated on the first aspect and have ‘graphical cadastres’, or ‘cadastral index maps’ that are compiled to cartographic or map standards rather than to survey measurement standards. In some countries the graphical cadastre is the only definition of cadastral boundaries. This is largely the case in England and Thailand, although in both countries there is the option to have specific boundaries defined by accurate ground survey. It is unrealistic to expect that a country or jurisdiction could undertake cadastral surveys of sufficient accuracy to solve all existing and future disputes over boundaries. Even in countries that have systems of accurate surveys, the survey measurements themselves do not solve all boundary disputes. In New South Wales where parcel boundaries are typically defined on the basis of accurate surveys, the courts have established the following hierarchy of evidence in re-establishing boundaries (Hallman 1973): 1. Natural features 2. Monuments 3. Old occupation, long undisputed 4. Abuttals 5. Statements of length, bearing or direction. C. Cost implications in designing a spatial framework Decisions on spatial frameworks for land records are typically made either explicitly or implicitly in the design of major programs for systematic registration. In reviewing international experience in strengthening land administration systems, Dale and McLaughlin (1999:46) provide the following indicative breakdown in costs where technical options can comprise a large percentage of the overall project costs: • institutional strengthening: 10-15 percent • mapping: 20-5 percent • adjudication and surveying: 30-50 percent • registration: 20-5 per cent. A review of international experience in 18 jurisdictions (Burns, 2007:94-95) determined that the unit costs for first registration in the countries reviewed range from about $10 to $55 per parcel. Pre-field costs – mainly the cost of geodetic control and base mapping – can be a significant cost as indicated in the cases of Moldova, Thailand, El Salvador and the rural project in Peru. The unit cost for pre-field activity in Thailand, mainly geodetic control, aerial photography and photo-mapping is relatively small due to the large number of titles projected in the third phase of the project (over 4.77 million titles). The unit cost in the field of boundary identification and surveying was a significant cost element in most projects (notably, Armenia, Kyrgyzstan, Moldova, El Salvador and the rural project in Peru). In Armenia significant cost saving for the survey activity were realised by contracting the activity to the private sector. No project in the developing world has been able to implement and sustain high-accuracy surveys over extensive areas of their jurisdiction. Those countries that have been successful in registering significant numbers of titles have tended to concentrate on relatively simple, low cost survey methods and produced graphical standard cadastral index maps. This was the approach in the urban project in Peru. There is limited information available on the cost/benefits of various technical options in a developing country. A recent report (Alemu2006) documents an investigation of eight technical options for one rural 101 village of 154 land parcels covering 120 hectares about 35 kilometres outside Addis Ababa in Ethiopia. The technical options tested were: • the use of hand-held GPS equipment to coordinate corner marks; • the traditional rope survey technique; • a combination of the traditional rope technique to determine parcel areas and hand-held GPS units to measure parcel centroids; • a tape-and-compass technique to produce sketch maps and determine parcel areas; • a combination of tape-and-compass surveys to determine parcel areas and hand-held GPS to map parcel centroids and corners; • surveys with electronic total stations to measure parcels corners and determine parcel areas; and • ortho-projected IKONAS high-resolution satellite imagery to photo-identify parcel corners and determine parcel areas. The economic life of the various items of equipment was estimated and the depreciated daily cost of the equipment was included in the cost analysis of the study, as was estimates for the salary costs of staff and other direct costs of the various methods. The results of the study are summarised in Table 8. Table 8 Summary of Cost and Time Estimates in Ethiopia (from Alemu 2006). Methodology Cost (US$) Survey time/speed (hours:minutes) /parcel /ha /parcel /ha Hand-held GPS 4.98 9.27 00:19 00:34 Rope only 0.81 1.50 00:15 00:28 Rope and hand-held GPS 0.97 1.81 00:17 00:30 Tape and Compass 18.18 33.66 01:34 02:53 Tape and Compass and hand-held GPS 18.29 33.80 01:36 03:00 Total Stations 7.27 13.54 00:23 00:44 IKONAS satellite imagery 14.23 26.52 00:17 00:31 The use of hand-held GPS equipment is relatively cheap and quick. However, significant capacity building is required for this equipment to be used by woreda staff. The use of tape and compass was the most expensive option, due to increased time in undertaking the surveys. The use of total stations was moderately expensive, largely due to the cost of equipment, and required significant capacity building. The use of high-resolution satellite imagery was very expensive, largely due to the cost of the ortho- projected imagery (equivalent to US$12.11/parcel). Recent experience suggests that ortho-rectified imagery can be procured at US$12-20/square kilometer (equivalent to US$0.12 to US$0.20/parcel for broad cover with an average parcel size of 0.5 ha), so this estimate is very high. The traditional rope survey method is clearly cheaper and requires no capacity building. This process however will not result in any cadastral maps and will provide limited information to settle any future disputes over boundaries. The recent experience in Ethiopia in second level certification has recently been reviewed (Orgut, 2010). The review noted that there were about 50 million parcels in the four main regions that needed to be covered by the second level certification process. This will require a significant systematic registration activity, even considering the extent of first level certification. The key findings of the review of recent experience were: • In Amhara technical assistance provided by Sida had facilitated the preparation of 1,612 second level certificates based on accurate surveys using total stations and RTK GPS equipment. Although the final report of the Sida consultants noted that it was possible for a survey team to complete 90-100 parcels per day, these figures were very high and were not substantiated by pilot 102 experience. The review found that the use of total stations and precise GPS equipment was expensive. • In the four main regions the USAID-funded ELTAP project had undertaken significant work with hand-held GPS equipment to survey and capture information for second level certification but this had resulted in less than 3,000 second level certificates being prepared in Oromia Region. The major concern with the use of hand-held GPS was the low accuracy (5-10 metres) but the review team felt that the technique was low cost and capable of producing cadastral index maps which may or may not be acceptable as an acceptable spatial framework for the second level certification process. • A pilot activity was undertaken in Fagita-Lekoma woreda in Amhara region to compile a cadastral index map using high-resolution satellite imagery (Quickbird imagery with a GSD of 0.62 metres) as a map base. The review found that the process was very participatory and that the land holders had quickly understood the imagery and could easily identify key features on the imagery. D. Recommendations for Ethiopia 1. Boundaries should be agreed and formalised in a manner that reflects local custom and practices with: • A preference in rural areas for general boundaries with boundaries publicly marked or occupied; • A process developed for boundary locations to be publicly agreed by the affected land holders in the kebele with the oversight of the Land Adjudication Committee; • The agreed boundaries charted on cadastral index maps; • Subsequently unclear or disputed boundaries investigated and agreed in a public process in the community that is oversighted by the kebele Land Adjudication Committee using the cadastral index maps as a reference only; • A process developed to address the situation where a boundary cannot be agreed in the kebele where expert evidence is presented to an appropriate forum to conclusively decide on boundary locations with parties to the dispute able to hire at their expense experts such as registered surveyors to present expert evidence at such forum; • Any survey plans or reports produced to be recorded, made publicly available and boundaries charted on the cadastral index maps; • Guidelines on boundaries and the above processes regarding boundaries prepared and widely disseminated. 2. A cadastral index map using an orthophotomap base (using either aerial photography or high resolution satellite imagery) should be agreed as the spatial framework for second level certification with the cadastral index maps unambiguously identifying the geographic location of and relationship between the land parcels. 3. Survey procedures should be developed to supplement the cadastral index maps where it is deemed necessary and the results of these surveys should to be recorded, made publicly available and boundaries charted on the cadastral index maps. These survey procedures are to be specified in a manner that ensures that the services of a surveyor are readily available to land holders and that the cost of surveys is not a barrier to formalisation. 4. Procedures should be developed to maintain and backup the cadastral index maps and associated survey records. 5. Steps should be taken to develop the infrastructure to prepare cadastral index maps and develop the spatial framework for land records. These steps should include: 103 • An investigation of the requirements for a zero order geodetic network in Ethiopia, a plan to complete a zero order network and the implications on the cadastral records; • The preparation of a cost effective strategy to acquire aerial photography and/or satellite imagery and produce the necessary base maps for the cadastral index maps; • The building of the necessary capacity to establish and maintain the spatial framework; • The establishment of a process to accredit registered surveyors in a manner that best serves the needs of society. 104 Attachment 6. Summary of the Final Business Processes and Data Management Study Since 1998, land administration activities in Ethiopia have resulted in the registration and certification of over 7 million landholdings.89 An additional 5 million holdings await certification. In some places, the holding rights are registered in the registry books, but the holding certificates have not been issued to the landholders. After the success that has been achieved, it seems that some stagnation has occurred. In addition to a backlog of certification, updating the registered information following transactions (inheritance, gift, and divorce) is not always formally recorded in the registry books. The four regions vary significantly in their updating procedures. Arguable, the success achieved was due to a “bottom-up� approach, in which the adjudication of land rights was done at the local (kebele and subkebele) levels. The subsequent focus of registration in the woredas may be a factor in the slowdown. Additionally, initiatives to improve the current situation have tended to focus on piloting cadastral surveys for second-level certification and the use of information technology, at the expense of improving the paper-based registration system, which is the norm in over 99 percent of the land administration offices at the woreda and kebele levels. The key findings and recommendations on improving the business process and data management are: a. Paper or digital? The current land registration system in Ethiopia is paper based. Some initiatives have been made to introduce computerization of the system. However, realistically, the majority of registration offices will have a paper-based system for a long time. Nevertheless, it makes sense to prepare for the eventual computerization of all offices. Changing to a uniform unique parcel identification number (UPIN), which does not include the holding number, is an important change that would facilitate computerization. Defining clear business processes to be followed consistently and written down is another improvement, which will help introduce computers. In general, it is much easier to introduce a computer system, if there is in place a well designed and well-running paper-based system. Many of the changes recommended in this report seek to achieve that. b. Data quality. Concerns have been expressed in most regions about the quality of some of the data in the registration system. Data quality can be assessed in terms of (a) accuracy (for example, land parcel area measurements), (b) up-to-dateness (that is, whether all transactions subsequent to 1st registration have been registered), and (c) completeness (for example, whether all legal holders are registered, such as wives). The authors recommend that a data audit is done as an integral part of the transfer of any operational registration responsibility from woredas to kebeles. Correcting inaccurate parcel areas is the biggest part of the task on a per-parcel basis. A sample test of parcel areas should be done to gain an estimate of the size of the task. c. Registry books. In a land registration system, the registry book is the primary record of legal rights in land. It should contain all of the essential information that shows who has what rights in which parcels of land. It also should be kept up to date with the changes in these details that result from transactions, such as inheritance, gift, exchange, divorce, and rent. In 2 of the 4 regions investigated, the design of the registry book does not meet these requirements. Amhara region’s book meets the requirements, and Tigray Region’s book partially meets them. The main shortcomings are a lack of space to detail all parcels in a holding, and even less space to record the changes resulting from transactions. The authors recommend that a new registry book design be introduced in Oromiya, SNNPR, and Tigray. The recommendation is for either a “two-pages- per-holding� book or a “page-per-parcel� book. If a parcel book is chosen, it should be linked to a “Holding Book� which will list all of the parcels in each holding. 89 This data is adapted from the executive summary (Orgut 2010e). 105 d. Book of holding certificates (BoHC). In Ethiopia, the “Book of Holding Certificates� (the “Green Book�) is the document given to each holder of land rights as evidence of which rights the holder has in which parcels of land. The contents of these books differ among the 4 regions. We recommend that the Green Book remain as a “Book of Holding Certificates� which is the case in 3 of the 4 regions, Tigray being the exception. We recommend that the contents of the Green Books be improved, to varying degrees, in all regions. The improvements should be implemented to produce a book format that contains all of the essential data. Regions may add to this “core data� according to local requirements. e. Document storage. The land administration system contains a very large quantity of legal documents. These include the registry books and “supporting documents.� The supporting documents are by far the larger in terms of volume. The storage facilities for these documents generally are very poor, in woreda and kebele offices. Files containing the documents often are on the floor or in heaps on desks or cupboards. There is a lack of shelves and lockable cupboards to protect the documents against fire, flood, rodents and theft. We recommend an improvement in the quantity and quality of storage facilities in woreda and kebele offices. f. Parcel numbering. Clear identification of each individual land parcel is one of the fundamental requirements of a land registration system; this is achieved by a UPIN. The four regions have different UPIN systems, if they have any system at all. The two systems used in Ethiopia are “parcel-based� and “holding-based.� We recommend that a genuine parcel-based UPIN is introduced in all four regions.90 An example of a parcel number in the proposed system is: AM/DD/35/239/012/2749. The six elements of this UPIN indicate region, zone, woreda, kebele, cadastral block, and parcel, respectively. Letters are proposed for indicating the region and zone, because they are more “user friendly� than numbers. The “cadastral block� element is inserted so that rural parcel numbers are compatible with urban parcel numbers. This UPIN format should be implemented nationwide, as it will make eventual computerization of land records much easier and ensure that rural parcels are integrated easily with urban parcels. g. Urban areas. In the larger urban municipalities visited (for example, Bole subcity in Addis Ababa), the impression is of a great deal of land office staff activity that is barely able to keep up with the work needed to be done. The formal land administration system in urban areas consists of 2 systems: (1) “permit�––basically a rent system in which landholders can be given notice at any time and (2) “lease�––in which lease agreements are made for fixed time periods, for example, 60 years. However, in cities and towns, there is a very large “informal� sector, within which people do not have any formally registered rights. In “periurban� areas (rural areas immediately around towns into which the town expands), there are problems of paying compensation to farmers whose lands are required for urban expansion. Part of the solution to this problem is to ensure that the registry books of rural kebeles are transferred to urban land administration offices when rural kebeles become urban kebeles. A common UPIN in urban and rural areas is essential to facilitate the transfer of land records from rural to urban areas and to enable more effective management of land. h. Kebele office operations. In the land administration system, the 13,846 kebele land offices generally play a secondary role to woreda offices. However, there are efforts in Amhara and Tigray to move responsibility for operational registration from woredas to kebeles. To implement the objective of having operational registry books in kebele offices, an ambitious and long-term 90 Such a system already exists in Oromia region. In Amhara region, the system is a holding identification number; SNNPR and Tigray do not have a system. 106 program for change is required.91 A simpler and cheaper way of improving the registration procedures is for kebele land offices to act as a “window.� Each registration application would be submitted and recorded in an application book at the kebele land office. The application documents then would be transferred by land office staff to the woreda office for processing. Completed registration documents then would be returned to the kebele office, where the kebele copy of the registry book would be completed. We recommend that the “kebele window� approach be implemented in regions that cannot afford to implement full registration operations at the kebele level. i. Woreda offices. Currently, the about 560 woreda land administration offices are the focus of land administration operational activities in the 4 regions. In the majority of cases in the 4 regions, registration (first-level and updating) takes place in the woreda office. Depending on the preferred solution of making kebele offices the locations in which registration takes place (point h. above) the role of the woreda office may need to change. If fully operational registration takes place in kebele offices, we recommend that woreda offices should be devoted to monitoring, auditing, training (of kebele offices and staff), and maintaining backups of the registration records held in the kebeles. j. Human resources and training. Human resources are essential for any administrative activities. Staff need to be (1) sufficient in number (but not too many), (2) trained, (3) motivated, and (4) prepared to stay in his/her job for a reasonable length of time. Clearly, some land administration offices have staff shortages. We recommend that there should be more initiatives to “train trainers,� so that one staff member from an office is formally trained, then can go back to his/her office and train other staff members. More training should be done to enable staff to “multi-task,� that is, to do several jobs within the office, which should lead to better job satisfaction and fewer resignations. Programs of major change, such as transferring operational registration from woredas to kebeles, should take account of the reality of staff numbers in planning how long the program will take to implement. Any training, whether formal or informal, needs to be supported with documentation clearly explaining how the work should be done. k. Business processes. Business processes are the detailed series of tasks that make up land administration activities such as first-level certification, second-level certification, and updating due to inheritance, gift, divorce, exchange, and rent. The main problem associated with these processes is not that they vary among regions but that they (1) are not clearly defined (that is, written down in detail using text and diagrams) and (2) often cannot be properly implemented due to shortcomings in documentation, in particular, badly designed registry books. These two problems need to be resolved so that business processes can be properly implemented. We recommend that each business process be defined by a clear “Business Process Diagram� (section 8, Orgut 2010e). Creating these diagrams will help land office staff and citizens to know what each process comprises. A number of changes are recommended to various parts of the registration procedures: 1) As land administration work in most kebeles does not constitute a full-time job, a full- time employee should be responsible for land administration in several (3–5) kebeles. 2) A fee should be charged for the issue of each Green Book. 91 The main elements of such program are to (1) assess the current situation in all of the 13,846 kebele offices, (2) calculate the investment needed in each office make it capable of supporting operational registration, (3) prioritize which kebele offices should become operational and when, (4) train suitable kebele staff (paid and voluntary) to do the work, (5) audit the existing registration data in registry books and Green Books to ensure that the data is accurate and up-to-date, (6) enter the audited data into the new registry books, and (7) continuously monitor and audit the registration activities in the kebele offices by woreda office staff. 107 3) Fees should be charged for transactions in all regions. 4) An arrangement should be made to take a photographer to the kebele center for an “en- mass� picture-taking exercise (because the lack of photographs is delaying the issue of Green Books). 5) A written notice should be posted on the kebele notice board about the documentary and other requirements for all types of transactions. 6) Every transaction should be publicized to make the process transparent and reduce fraud. 7) When a client applies for a service, the application should be recorded in an application journal and the client should sign the journal when s/he receives the updated green/yellow book. 8) Rental agreements for longer than three years should be registered. l. Updating (transactions). The requirement for updating begins as soon as initial registration is done. Changes in land rights, land parcels, and landholdings occur daily due to inheritance, divorce, rent, gift, exchange of parcels, parcel subdivision, and parcel amalgamation. Although registering these changes does occur in all four regions, there is much variation in how thoroughly and accurately registration is done. The inconsistencies arise from (a) badly designed registry books; (b) lack of training on the importance of updating; (c) lack of clear written procedures for updating; and (d) informal transactions among land rights holders, which is due to a lack of public awareness of, and the inconvenience of having to go to, the woreda office to register transactions. A public awareness campaign is particularly important measure to reduce the number of informal transactions is reduced. 108 REFERENCES Adal, Yigremew. 2005. Rural Land Certification in Ethiopia. Working Paper. Institute of Development Research, Addis Ababa. Alemu, D. 2006. Comparative Performance Assessment of Various Cadastral Surveying Methodologies: Case- study of the Hordofi Village Cadastral Surveying Exercise. Unpublished report prepared by ARD Inc. in collaboration with LALU and LAUD. April. Andersen, Kirsten Ewers, and Julien Dupuy. 2009. Study on the Identification and Formation of User Groups for Mapping, Registration, and Management of Communal Lands and Common Pool Resources, Ethiopia. Final Version. FAO, Rome. December. Anteneh, Addis. 2007. The Assessment of Rural Land Valuation and Compensation Practices in Ethiopia. Final Main Report. Ethiopia Land Tenure and Administration Program (ELTAP), Addis Ababa. Anteneh, Addis. 2008. Pastoral and Agro-Pastoral Land Tenure and Administration Study. Main Report. Final. Ethiopia Land Tenure and Administration Program (ELTAP), Addis Ababa. Ayano, Mekonnen F. 2009. Agrarian Law and Development in Ethiopia. Submitted Dissertation for LL.M degree, Harvard University, Cambridge, MA. Ballaney, Shirley. 2010. Town Planning Scheme Mechanism Gujarat. Annual Bank Conference on Land Policy and Administration, 26 – 28 April 2010, Washington DC, International Workshop on Expanding Urban Land Supply: What Works? http://siteresources.worldbank.org/INTHOUSINGLAND/Resources/339552- 1272658002920/Ballaney.pdf BPN (Badan Pertanahan Nasional). 1993. Land Administration Project. Project Preparation Report. Unpublished. May. Bowman, Chakriya. 2004. Thailand Land Titling Project. Paper presented at Conference on Scaling up Poverty Reduction, Shanghai, May 25–27. http://info.worldbank.org/etools/docs/reducingpoverty/case/101/ fullcase/Thai%20Land%20Titling%20Full%20Case.pdf Brits, Anne-Marie, Christopher Grant, and Anthony Burns. 2002. Comparative Study of Land Administration Systems with Special Reference to Thailand, Indonesia and Karnataka, India. Synthesis paper, World Bank. May. Bruce, John, and Zongmin Li. 2009. Crossing the River While Feeling The Rocks: Incremental Land Reform and Its Impact on Rural Welfare in China. IFPRI (International Food Policy Research Institute) Discussion Paper 926. Washington, DC. Burns, Anthony. 2007. Land Administration Reform: Indicators of Success and Future Challenges. ARD (Agriculture and Rural Development) Discussion Paper 37. World Bank. http://siteresources.worldbank.org/ EXTARD/Resources/ARDDiscussionPaper37.pdf Dale, P.F., and John D. McLaughlin. 1988. Land Information Management: An Introduction with Special Reference to Cadastral Problems in Third-World Countries. New York: Oxford University Press. _____. 1999. Land Administration. New York: Oxford University Press. Deininger, Klaus. 2003. Land Policies for Growth and Poverty Reduction. A World Bank Policy Research Report. Oxford and New York: World Bank and Oxford University Press. Deininger, Klaus, and Gershon Feder. 2009. Land Registration, Governance, and Economic Development: Evidence and Implications for Policy. World Bank Research Observer 24 (2): 233–66. Deininger, Klaus, and Songqing Jin. 2006. Tenure Security and Land-Related Investment: Evidence from Ethiopia. European Economic Review 50 (5):1245–77. _____. 2008. Land Sales and Rental Markets in Transition: Evidence from Rural Vietnam. Oxford Bulletin of Economics and Statistics 70 (1): 67–101. 109 _____. 2009. Land Rental Markets in the Process of Rural Structural Transformation: Productivity and Equity Impacts from China. Journal of Comparative Economics 37 (4): 548–66. Deininger, Klaus, Daniel Ayalew Ali, and Tekie Alemu. 2011. Impacts of Land Certification on Tenure Security, Investment, and Land Market Participation: Evidence from Ethiopia. Land Economics 87(2): 312-334. Deininger, Klaus, Daniel Ayalew Ali, Stein Holden, and Jaap Zevenbergen. 2008. Rural Land Certification in Ethiopia: Process, Initial Impact, and Implications for Other African Countries. World Development 36 (10): 1786–812. DHV Consultants. 2006a. Integrated Urban Land Information Systems: Inception Report. Ministry of Federal Affairs, Federal Democratic Republic of Ethiopia. Addis Ababa. _____. 2006b. Integrated Urban Land Information Systems: Implementation and Final Report. Ministry of Works and Urban Development, Federal Democratic Republic of Ethiopia. Addis Ababa. Do, Quy-Toan, and Lakshmi Iyer. 2008. Land Titling and Rural Transition in Vietnam. Economic Development and Cultural Change 56 (3): 531–79. Field, Erica. 2005. Property Rights and Investment in Urban Slums. Journal of the European Economic Association 3 (2–3): 279-90. FAO (Food and Agriculture Organization of the United Nations). 1986. Highlands Reclamation Study: Ethiopia. Final Report. Two vols. Rome. _____. 1996. Effective Control of Urban Development by Land-Use Plans. Rome. Federal Democratic Republic of Ethiopia. 2001. The Government of the Federal Democratic Republic of Ethiopia, Rural Development Policies, Strategies and Instruments. Press and Audiovisual Department, Ministry of Information (MoI). Addis Ababa. November. _____. 2003. Government of the Federal Democratic Republic of Ethiopia Rural Development Policy and Strategies. Ministry of Finance and Economic Development (MoFED). Addis Ababa. _____. 2008. Ethiopian Sustainable Land Management Investment Framework (ESIF). Ministry of Agriculture and Rural Development (MoARD). Addis Ababa. Fransen, Jan, and Meine Pieter van Dijk. 2008. Informality in Addis Ababa, Ethiopia. Paper prepared for the conference, Are Cities More Important than Countries?’ Erasmus University, Rotterdam, October 30. www2.econ.uu.nl/users/marrewijk/pdf/ihs%20workshop/ fransen%20paper.pdf Galiani, Sebastian, and Ernesto Schargrodsky. 2005. Property Rights for the Poor: Effects of Land Titling. Working Paper 06/2005. Universidad Torcuato Di Tella, Escuela de Negocios, Centro de Investigación en Finanzas, Buenos Aires. http://cdi.mecon.gov.ar/biblio/ doc/utdt/cif/wp2005-06.pdf GTZ-IS. 2006. Urban Land Administration and Land Markets. Final Report for the Ministry of Works and Urban Development, Urban Development Capacity Building Office, Federal Democratic Republic of Ethiopia, Addis Ababa. June. Haile, Mitiku, Wray Witten, Kinfe Abraha, Sintayo Fissha, Adane Kebede, Getahun Kassa, and Getachew Reda. 2005. Land Registration in Tigray. Research Report 2. Mekelle University, Tigray. www.eldis.org/vfile/upload/1/document/0708/ DOC20756.pdf Hallman, Frank. 1973. Legal Aspects of Boundary Surveying as apply in New South Wales, Published by the Institution of Surveyors, Australia, New South Wales Division. Heurlin, Christopher. 2007. China’s New Rural Insurgency: The Politics of land Expropriation. Paper presented at the Annual Meeting of the International Studies Association, 48th Convention. Chicago. http://www.allacademic.com/meta/ p180992_index.html Holden, Stein, and Tewodros Tefera. 2008. From Being Property of Men to Becoming Equal Owners? Early Impacts of Land Registration and Certification on Women in Southern Ethiopia. Research Report submitted to UN-HABITAT, Shelter Branch, Land Tenure and Property Administration Section. Nairobi. 110 http://www.gltn.net/en/e-library/gender/from-being-property-of-men-to-becoming-equal-owners-early- impacts-of-land-registration-and-certification-on-women-in-southern-ethiopia/details.html Holden, Stein, Klaus Deininger, and Hosaena Hagos Ghebru. 2009. Impacts of Low-Cost Land Certification on Investment and Productivity. American Journal of Agricultural Economics 91 (2): 359–73. _____. 2011. Tenure Insecurity, Gender, Low-Cost Land Certification and Land Rental Market Participation. Journal of Development Studies 47(1), 31-47. Imeru, Tamrat. 2009. Improving the Land Governance Sector in Ethiopia. Consultant Report. Multi-Talent Consultancy Pl, Addis Ababa. Kaganova, Olga. 2005. Establishing More Efficient and Effective Land Management in the Amhara and Tigray Regions, Ethiopia. Final Report prepared for Capacity Building for Decentralized Service Delivery Project, Urban Development Capacity Building Office, Ministry of Federal Affairs, Federal Democratic Republic of Ethiopia. Addis Ababa. Kaganova, Olga. 2006. Land Delivery Mechanism, Land Expropriation, and Property Valuation. Urban Institute, Washington, DC. March. Kirk, Michael, and Nguyen Do Anh Tuan. 2009. Land-Tenure Policy Reforms: Decollectivization and the Doi Moi System in Vietnam. IFPRI (International Food Policy Research Institute) Discussion Paper 926. Washington, DC. Liu, Guo-zhen. 2007. The Reform of the Compensation System of Land Expropriation in China. Journal of US- China Public Administration 4 (1):11–20. www.managers.org.cn/mag/doc/ucman200701/ ucman20060702.pdf MoARD (Ministry of Agriculture and Rural Development). 2009. Plan for Building Sustainable Land Administration and Use System at the National Level. Federal Democratic Republic of Ethiopia, Addis Ababa. In Amharic. _____. 2010. Concept Note on the Design and Implementation of Ethiopia: Land Administration and Land Use Development Project (ELALUDEP). Federal Democratic Republic of Ethiopia, Addis Ababa. April. MoFED (Ministry of Finance and Economic Development). 2006. Ethiopia: Building on Progress. A Plan for Accelerated and Sustained Development to End Poverty (PASDEP) 2005/06–2009/10. Federal Democratic Republic of Ethiopia, Addis Ababa. _____. 2010. The Federal Democratic Republic of Ethiopia: Growth and Transformation Plan. Draft. Federal Democratic Republic of Ethiopia, Addis Ababa. September. Ministry of Lands and Forestry. 2002. Project Preparation Report: Ghana Land Administration Project: Design and Implementation Plan. Republic of Ghana, Accra. March. Orgut Consulting AB. 2010a. Study on Institutional Structure and Human Resource Development Needs in the Sector of Land Administration in Ethiopia. Final Report for World Bank, Ethiopia. Stockholm. January. _____. 2010b. Specification of Spatial Framework and Detailed Implementation Strategy. Final Report for World Bank, Ethiopia. Stockholm. June. _____. 2010c. ICT Strategy for Land Record. Final Report for World Bank, Ethiopia. Stockholm. June. _____. 2010d. Software Strategy for Land Records Version. Final Report for World Bank, Ethiopia. Stockholm. June. _____. 2010e. Business Process Data Management Strategy. Final Report for World Bank, Ethiopia. Stockholm. June. _____. 2010f. Development of MIS, Performance Monitoring, Report, and Resource Planning. Final Report for World Bank, Ethiopia. Stockholm. June. 111 Pingali, Prabhu, and Vo Tong Xuan. 1992. Vietnam: De-collectivization and Rice Productivity Growth. Economic Development and Cultural Change 40 (4): 697–718. Rahmato, Dessalegn. 2007. Land Rights and Tenure Security: Rural Land Registration in Ethiopia. Occasional Paper 2, Van Vollenhoven Institute for Law, Governance and Development, Faculty of Law, Leiden University, Netherlands. Rattanabirabongse V., Robert A. Eddington, Anthony Burns and Kevin G. Nettle. 1998. The Thailand Land Titling Project: Thirteen Years of Experience. Journal of International Land Use Policy 15 (1): 3–23. Regional Government of Amhara in collaboration with World Bank and UN-HABITAT GLTN. 2009a. Application of High-Resolution Satellite Image for the Production of Cadastral Index Mapping (CIM). Bahir Dar. _____. 2009b. The Possibility of Using a Pro-Poor, Open Source Technology (STDM) for Land Administration and Record Management. Bahir Dar. Ryskamp, John. 2006. The Eminent Domain Revolt: Changing Perceptions in the New Constitutional Epoch. New York: Algora Publishing. Sommer, Eric. 2010. Amend Law to Prevent Forced Demolitions. China Daily. June 3. http://www.chinadaily.com.cn/opinion/2010-04/27/content_9783079.htm Teklu, Askale. 2005. Land Registration and Women’s Land Rights in Amhara Region, Ethiopia. Securing Land Rights in Africa Research Report 4. International Institute for Environment and Development (IIED), London and Edinburgh. November. http://www.oxfam.org.uk/resources/ learning/landrights/downloads/land_reg_womens_land_rights_amhara_iied_res_rep4.pdf USAID (United States Agency for International Development). 2008. Ethiopia: Strengthening Land Tenure and Administration Program (ELTAP): Evaluation. Final Report. Washington, DC. April. Williamson, Ian. 2000. Best Practices for Land Administration Systems in Developing Countries. Paper presented at the International Conference on Land Policy Reform, Jakarta. July. http://siteresources.worldbank.org/INTARD/825826-1111055015956/20424575/ lapcWilliamson.pdf Woldemariam, Asmelash. 2006. The Effects of Land Reform on Peasant Social Organization: A Study of Village- Level Dynamics in Central Tigray, 1974–1994. Anthropology Dissertation Series 7. Department of Sociology and Social Anthropology, Addis Ababa University. World Bank. 2006. Poverty and Land Degradation in Ethiopia: How to Reverse the Spiral? Environment, Rural and Social Development Operations, AFTS2, Africa Region. _____. 2007. The Challenge of Urbanization in Ethiopia: Implications for Growth and Poverty Alleviation. Water and Urban Development Unit 1, Africa Region. _____. 2008. Project Appraisal Document: Federal Democratic Republic of Ethiopia: Sustainable Land Management Project. Environment and Natural Resources Management, Sustainable Development Department, AFTSDN, Africa Region. _____. 2009a. Bulgaria: Registration and Cadastre Project Implementation Completion Report. ICR1138. _____. 2009b. Toward the Competitive Frontier: Strategies for Improving Ethiopia’s Investment Climate. Report No. 48472-ET. Africa Region. 112