Report No. 38199-GE Georgia Rural Infrastructure in Georgia Improving Service Delivery (In Two Volumes) Volume I: Main Report December 14, 2006 Transport Unit, Sustainable Development Department Europe and Central Asia Region Document of the World Bank Rural Infrastructure in Georgia ­ Improving Service Delivery TABLE OF CONTENTS Acknowledgments ............................................................................................................................................... iii The Development of Rural Infrastructure in Georgia...................................................................................... 1 A Coherent Rural Infrastructure Strategy for Georgia.................................................................................... 1 What Are the Key Issues That a Rural Infrastructure Strategy Should Address?........................................... 3 What Can Be Done to Make Rural Infrastructure Investments Sustainable? ................................................. 7 The Output From the Decision Matrix.......................................................................................................... 14 How Much Would the Investment Plan Cost and How Would It Be Financed? .......................................... 16 Key Recommendations................................................................................................................................. 19 The Findings of the Study ................................................................................................................................. 21 What Is the Coverage and Condition of Rural Infrastructure?...................................................................... 21 What Are the Community Preferences in Rural Infrastructure?................................................................... 40 What Is Household Willingness to Pay for Rural Infrastructure?................................................................. 46 What Is Estimated Cost to Rehabilitate the Rural Infrastructure Networks?................................................ 52 Key Conclusions and Next Steps....................................................................................................................... 61 References........................................................................................................................................................... 63 Appendix 1 ­ The Decision Matrix................................................................................................................... 66 Appendix 2 ­ An Example of the Output From the Decision Matrix............................................................ 68 LIST OF FIGURES Figure 1 : Sectoral Composition of the Highest Scoring Interventions................................................................ 14 Figure 2 : Sectoral Composition of the Interventions with Highest Emphasis on Poverty Reduction................. 15 Figure 3 : Sectoral Composition of the Interventions with Highest Emphasis on Economic Development........ 15 Figure 4 : Hours of electricity supply per day by rayon (2005)........................................................................... 22 Figure 5 : Sakrebulos with Natural Gas Supply (2005) ....................................................................................... 27 Figure 6 : Schools with Water Supply by Sakrebulo (2005)................................................................................ 30 Figure 7 : Water and Sewage Infrastructure Condition (2005)............................................................................ 32 Figure 8 : Landline Telephone Coverage (2005) ................................................................................................. 35 Figure 9 : Cellular Telephone Coverage (2005) .................................................................................................. 36 Figure 10 : Road Network and Condition in Georgia (2005)............................................................................... 38 Figure 11 : Example of Community Meeting and Focus Group conducted during the Study ............................. 40 i LIST OF TABLES Table 1 : Access, availability and consumption of electricity within households, by region............................... 23 Table 2 : Supply of electricity to communities by electricity provider (% of communities) ............................... 24 Table 3 : Condition of electricity network (% of communities) .......................................................................... 25 Table 4 : Access to Gas Supply System............................................................................................................... 26 Table 5 : Condition of Gas Supply System by community and region................................................................ 28 Table 6 : Access to water and main source of water by region (% of households, % of communities)............... 29 Table 7 : Condition of landline network (% of communities) ............................................................................. 33 Table 8 : Comparative Road Indicators by Region.............................................................................................. 37 Table 9 : Percentage of irrigable land receiving irrigation in communities (units: % of surveyed communities) 39 Table 10 : Condition of irrigation systems (units: % of surveyed communities)................................................. 40 Table 11 : Proportion of communities expressed preference for infrastructure rehabilitation, by type............... 41 Table 12 : Expressed Preferences towards Community Ownership .................................................................... 46 Table 13 : Willingness to pay estimates for service improvements, by expenditure group, GEL per month ...... 48 Table 14 : Willingness to pay estimates for road improvements, by expenditure group, GEL per month (according to the estimated model)............................................................................................................... 50 Table 15 : Sum of willingness to pay estimates, GEL per month (according to the estimated model)................ 50 Table 16 : Willingness to pay estimates for Package 1 and Package 2, GEL per month (according to the estimated model)........................................................................................................................................... 51 Table 17 : Relation between the sum of WTP and maximum WTP for all services (according to the estimated models) ......................................................................................................................................................... 51 Table 18 : Road interventions and unit costs ....................................................................................................... 52 Table 19 : Road Investment Cost by region......................................................................................................... 53 Table 20 : Water supply investment components and unit costs ......................................................................... 54 Table 21 : Estimated Unit Costs for Installation of New Gas Infrastructure ....................................................... 55 Table 22 : Gas distribution rehabilitation costs by region ................................................................................... 55 Table 23 : Electric distribution network rehabilitation costs by region ............................................................... 57 Table 24 : Drainage and irrigation investment needs by regions (US$ Millions)................................................ 58 Table 25 : Input variables to estimate telecommunications investment costs...................................................... 59 Table 26 : Telecommunication investment needs by region................................................................................ 60 ii ACKNOWLEDGMENTS This paper was prepared by a team led initially by Gevorg Sargsyan (Sr. Infrastructure Specialist, ECSSD) and later by Richard Martin Humphreys (Sr. Transport Economist, ECSSD), and finalized under the supervision of Olivier Le Ber (Sr. Transport Specialist, ECSSD). The team comprised the following members, in alphabetical order, Satoshi Ishihara (Social Development Specialist, ECSSD), Sevara Melibaeva (JPA, ECSSD), Joseph Melitauri (Project Officer, ECSSD), Julie Morel (Short Term Consultant, ECSSD), Maria Shkaratan (Short Term Consultant, ECSSD), Sanjay Sinha (Operations Officer, ECSSD), and Tamara Sulukhia (Senior Infrastructure Specialist, ECSSD) from the Europe and Central Asia Region of the World Bank. The external contributors were Mikkel Birkelund (COWI Consulting Group) and his team, who undertook the willingness to pay study; Dr. Anna Alberini (University of Maryland), who fulfilled a quality assurance role for the willingness to pay study; Zurab Jincharadze, who undertook the road condition survey; and Gocha Tskitishvili (Institute of Polling and Marketing) and his team, who undertook the quantitative and qualitative survey of the local communities. Thanks are also given to Peter D. Thomson (Director, ECSSD), Motoo Konishi (Sector Manager, Transport, ECSSD), Joop Stoutjesdijk (Lead Irrigation Engineer, ECSSD), Christian Petersen (Lead Economist, ECSPE), and Rosalinda Quintanilla (Lead Economist, ECSPE) for their comments on the draft of this paper. The team would also like to thank the Peer Reviewers Maninder Gill (Sector Manager, ECSSD), and John Hine (Sr. Rural Transport Specialist, TUDTR), for their helpful and substantive contributions to the draft of the paper. The team would also like to take this opportunity to acknowledge the generous contribution provided by a number of representatives of the Government of Georgia over the duration of this study. iii THE DEVELOPMENT OF RURAL INFRASTRUCTURE IN GEORGIA 1. Rural infrastructure in Georgia is at present in a precarious state and is not providing the services needed for the socio-economic development of the regions. Fifty percent of the labor force is currently engaged in rural activities and poverty is widespread. A broad strategy and detailed action plan are needed for the development of the physical rural infrastructure in Georgia. The intention of this study's findings is to enhance a wider dialogue between the Government, key stakeholders, and the general public on the development of a prospective Rural Infrastructure Strategy. The primary emphasis in this study is on physical infrastructure: roads, drinking water, irrigation and drainage, electricity supply and gas supply. This focus is in accordance with the wishes of the Government of Georgia, earlier research conducted by the World Bank1, and the feedback during the design stage of the study. An additional benefit was the opportunity to focus the study on those infrastructure types that had similar network and operational characteristics. 2. The study has conducted a range of surveys across Georgia in order to address a number of issues: (i) What is the coverage and condition of rural infrastructure? (ii) What are the community preferences in rural infrastructure? (iii) What is the household Willingness to Pay (WTP) for rural infrastructure? (iv) What is the estimated cost to rehabilitate the rural infrastructure networks? A more detailed description of the study approach, the sampling methodology, and the findings are provided in the relevant Annexes in Volume II2. A COHERENT RURAL INFRASTRUCTURE STRATEGY FOR GEORGIA 3. Any serious attempt to address the issues of the rural sector in Georgia will require the development of a comprehensive and coherent strategy, together with an action plan, for a program of rural infrastructure rehabilitation. It should specify the types of infrastructure, whether basic needs provision is sufficient, or whether more expensive technical designs are appropriate to try and facilitate broader economic development, and in which locations, all within a long term strategy for the development of rural communities in Georgia. The 1See World Bank (2001) for a discussion of the priorities of the poor and non-poor in respect of infrastructure types. 2 Volume II of this report along with the Decision Matrix are available on the CD ROM and website: http://lnweb18.worldbank.org/ECA/Transport.nsf/Countries/Georgia?Opendocument 1 development of strategy and action plan should reflect broad based consultation with key stakeholders and the general public. 4. Within such a strategy and action plan, the institutional and operational ownership of the infrastructure would need to be clarified to provide appropriate maintenance, ensuring that where a private, or public, sector organization is endowed with the responsibility for service provision and/or maintenance, the incentives are sufficient to ensure that the task is fulfilled and sustained. This implies that when a particular rural infrastructure intervention3 is designed for the community, consideration needs to be given to its fit within the national strategy framework, the appropriate institutional framework for service delivery, and local needs and resources, to ensure that it will be sustainable. 5. The Rural Infrastructure Development Strategy and action plan defined in this study would also need to be accompanied by an expenditure plan, consistent with the Government of Georgia's Basic Data and Directions (BDD) document, prepared in 2006, and the wider macroeconomic framework. The expenditure plan would need to clearly indicate the respective portions of the necessary financing to be provided by the Government of Georgia and by external donors, and the proportion, and type, of contribution required from the communities. The strategy would also need to consider the following points in the definition: x The relationships between rural infrastructure and current/potential agricultural growth including non-farm growth. x Possible delineation of rural communities into communities with low potential for growth and therefore with basic infrastructure requirements to provide their subsistence needs vis-à-vis those communities with high economic development potential, as a way of defining the appropriate design and cost of particular interventions, and focusing limited resources in those areas where it is most deserved. x Maximizing complementarity between different infrastructure services, such as roads and irrigation, to maximize benefits. Often little effect can be achieved by increasing crop yield, if the increase would have to be dumped due to poor roads. x Linkages with other rural development issues: rural infrastructure is necessary but often not sufficient for facilitating the agricultural growth and increasing market access. Other issues such as increasing productivity, making financing available, providing support for marketing, and building capacity should also be addressed. Thus, the rural infrastructure interventions, beyond addressing basic needs of population, should be synchronized with and often preceded by above mentioned interventions. 3For definition of different types of rural infrastructure interventions see Chapter on Findings of the Study. 2 The need for adequate monitoring of strategy implementation and infrastructure condition 6. The implementation of the strategy and the action plan and the condition of rural infrastructure services should be periodically monitored. It would appear prudent that, if the Government decides to follow this recommendation, then suitable questions are included in the annual Household Survey, undertaken by the State Department of Statistics of Georgia (SDSG), to assess respondent perception of infrastructure quality in their communities. This could be done at relatively little cost, which could then be incorporated into a broader, more systematic, assessment of strategy implementation. The tools and methods developed and applied during this study can be used for that purpose, with the results providing a baseline against which improvement could be measured. WHAT ARE THE KEY ISSUES THAT A RURAL INFRASTRUCTURE STRATEGY SHOULD ADDRESS? 7. There is a need for greater co-ordination between donors under Government ownership. The recommendation is that the Government should exercise greater ownership of the different initiatives within the sector, with the agreement of all stakeholders, and enhance the co-ordination between the donor community, within the context of a defined sector strategy, action plan, and expenditure plan for the rural infrastructure sector. The need for greater donor co-ordination underlines the importance of the development of instruments to guide all interventions in the sector. 8. There is a clear need for one of the existing ministries/agencies to take a central role in the definition, implementation and monitoring of the strategy, as well as take a leading role in the sector coordination of donor-funded activities. In recent years, the Government of Georgia has been improving the donor coordination under the leadership of the Ministry of Finance. The implementation of municipal, social and irrigation projects has been mobilized within one institution ­ Municipal Development Fund of Georgia (MDF) ­ in order to improve the coordination of infrastructure and development projects.4 Substantial progress in this direction has been achieved and should be continued. What is an appropriate level of access for the Government to provide? 9. One main finding from this study is that rural residents want inexpensive, reliable, access to rural infrastructure. This implies a policy trade-off between access and quality, which was made explicitly by a number of respondents. The issue, and there are significant cost implications in the decision, is what the appropriate scale and scope of access should be in the different sectors, and across the communities. A Minimum Level of Service (MLS) should be defined for each rural infrastructure type, which would ensure a basic access to the specific infrastructure, to minimize poverty in its broadest sense. In those communities where the potential for economic development is more significant, the level of provision could be much higher, but would be defined by the communities themselves, based on the 4In January 2006, Georgian Social Investment Fund (GSIF) was integrated into MDF, and in August 2006, World Bank financed Irrigation and Drainage Community Development Project was moved from the Ministry of Agriculture to the MDF. 3 willingness to pay of inhabitants and their specific needs. These issues are key policy questions to be considered in the development of the strategy. 10. There is no clearly formulated policy regarding the provision of a MLS for rural infrastructure, with one exception, the power sector. This is the only sector where the concept has been specified in the legislation. But what are the alternative choices in respect of the provision of MLS that the Government might want to consider for a strategy? Different strategies have different cost implications and they need to balance provision of MLS with alternatives, based on cost implications as well as other considerations. A number of different strategies were identified during the course of the study, although it was not possible to include all of them in the Decision Matrix. What should be done to improve the situation? 11. Define the concept of MLS in the legislation for all sectors - the MLS concept has two main components, namely provision of MLS and tariff. The MLS concept should not necessarily mean private access to all services. In the case of telecom, for example, the majority of people in the rural communities would prefer to have reliable payphones, rather than individual telephones in each household. An analogy can be drawn with water supply; while people might prefer to have individual taps, given the cost differential, they are likely to be satisfied with good quality water from boreholes or standpipes. The key issue is, rather than impose a solution on the rural communities, which may be non-optimal and therefore unlikely to be sustainable, it is important to allow the people to choose the quality and price of services they can afford, above a basic minimum level of service. 12. Develop a mechanism to finance a MLS - There appear to be two possible options: x Direct subsidies from the state budget to the operator targeted at providing access to a particular group of beneficiaries. This approach is more transparent and might be more efficient especially if there are competing service providers. The problems are scarce public funds and also targeting. Although, the latter might be a problem in the case of other schemes as well. The donor funding might be used to supplement scarce public resources. One way around these problems is for an explicit Public Service Contract, drawn up between the Government and a private sector provider, which specifies the required levels of service provision and quality, and the resulting compensation. Contracts such as these are used to deliver socially necessary train services in EU countries. x Establishment of shared financing mechanism for MLS, whereby the service providers are required to contribute defined amounts, in the form of a levy, which can then be used to pay for the service of remote areas and poor residents. This mechanism needs to be defined within the budget and subject to all fiduciary rules of the budget. This is a form of cross-subsidy; however, it does not have some short-comings of implicit cross-subsidization of lacking transparency and accountability. Namely, it is transparent and by being defined within the budget it is subject to all fiduciary rules of accountability. In addition, it aims at encouraging investments in remote areas as well as for the application in an 4 environment with multiple service providers. The draft electronic communication law proposes to establish such a mechanism for providing the MLS for fixed line telephones. 13. The overall recommendation is to go for prices at cost recovery with explicit subsidies to increase access in poor underserved areas. The subsidies need to be determined based on the results rather than earmarking budget resources. One possibility is to determine the unit cost of service and allocate budget resources for the subsidy based on the number of households served. So long as the service is metered, it may be more expensive but people will pay for it because they can manage their individual expenditures better. 14. In regard to road infrastructure, we recommend a phased approach to the implementation of the 2005 Law on Local Self Government (LSG) Units to avoid extending the current unfulfilled/unfunded mandate. The new 2005 Law, that became effective in November 2006, states that all local roads are now the responsibility of the relevant LSG. As a result the length of roads under each LSG will increase substantially. This approach has been adopted in many countries and has in many cases resulted in both an unfunded or unfulfilled mandate, in that sufficient resources and capacities are not available to the LSGs to fulfill their responsibilities. It would be prudent to introduce this approach in a phased manner in one, or two, pilot regions to allow problems to be identified and addressed, before wider implementation. It is also recommended that the Road Administration remains involved (in a way to be determined) in assisting the LSGs in managing the local roads transferred under their responsibilities. What could be done to improve the quality of service delivery? 15. There is a need to introduce new technologies and standards - One concomitant to the definition of minimum level of service for each infrastructure service is that the government needs to revise, where necessary, current norms and standards and introduce new technologies to meet the requirements implied by the interpretation. There is a clear trade- off between service quality and cost, which needs to be clearly reflected in the strategy for the sector, allowing more cost effective technologies to be employed in those communities with low potential for growth. 16. As one specific example, the geometric standards and technical specifications in the road sector need to be amended to reflect current conditions. A further unresolved issue pertains to the geometric standards and technical specifications which are used on the roads in Georgia at present. The existing standards in the road sector, inherited from the Soviet Union, for the rehabilitation and maintenance of rural roads appear excessive, when considering the traffic volumes on these roads. Current technologies are too expensive and more appropriate technologies such as concrete or bitumen stabilization, using bitumen emulsion, or slurry seal, could be used in many cases. As a result, roads are sometimes constructed to standards in excess of need and maintenance capacity. The introduction of 4.5 meters roads in mountainous areas, or gravel roads, with no shoulders, and appropriate new technologies could realize unit cost savings of 40-50 percent in rehabilitation activities. The WB financed Secondary and Local Roads Projects aims to assist the Government in designing and adopting new geometric standards for roads. The introduction of appropriate design standards for rural roads is a priority. 5 17. A general recommendation is that the defined strategy and policy should seek to maximize the use of new technologies and more appropriate standards that are likely to reduce capital and maintenance cost and/or improve quality of rural infrastructure services. This will have two main benefits; firstly, it will allow a better use of limited resources, thereby allowing a scaling up of interventions for a given budget; and secondly, it will enhance the sustainability of the interventions by more closely linking identified need with design. 18. There is a need to maximize the contribution of private sector in service provision in rural infrastructure. The study has revealed that there are opportunities for private sector contributions in the provision of rural infrastructure services. Some opportunities have been already utilized, including the contracting out of construction, rehabilitation and, to a certain degree, the maintenance of some physical infrastructure (e.g. some local roads). There are a number of other areas, where private sector involvement should be considered: (i) outsourcing management of community owned utilities, particularly drinking water; (ii) outsourcing operation, including maintenance, billing and collection to irrigation Water User Associations; (iii) designing and implementing credit (micro) financing schemes for communities as well as individual community members to improve infrastructure services, e.g. gas, water, communications; and (iv) cooperation in providing MLS as described above. The capital costs of access to infrastructure services should be funded by the private sector to the extent that these costs are financially viable. 19. The WTP survey has revealed considerable WTP on the part of rural inhabitants (see Table 16), with one example: households with an income of less than GEL 100 per month, that currently purchase water from vendors, reported a WTP of GEL 25 per month to receive piped water into the dwelling. The necessary capital for the connection may still be problematic for such households, and micro financing facilities could be made available to try and overcome this impediment. 20. Central and local governments should bear the access costs for those communities and households, where the marginal costs of connection outweigh marginal revenues, or ability to pay, and where connection is considered socially desirable. The appropriate level of access funded through the government funds should be determined based on the MLS concept. 21. There is a need to encourage public-private partnerships in funding capital investments for the installation of internal transmission network, and specifically the Government should encourage donors' engagement in Community Driven Development (CDD), including Georgian Social Investment Fund (GSIF)5, and rural development activities to consider funding this portion of capital costs for the private infrastructure service providers (particularly for the telephone and gas companies). 5 In January 2006, GSIF was integrated into the MDF as part of Government's effort to concentrate the implementation of municipal, social and irrigation projects within one institution. 6 WHAT CAN BE DONE TO MAKE RURAL INFRASTRUCTURE INVESTMENTS SUSTAINABLE? How to encourage people to pay? 22. The billing and metering systems need to be improved and managed efficiently, as they create disincentives to pay as many of the household customers decide not to pay for poor quality services or when they expect to be disconnected for non-payment by others. The farmers pay for irrigation by hectares or by hours. The insufficient water flow or the inaccurate calculations cause significant disparities between water charges and usage, leading to widespread criticism. 23. A further problem is that farmers are, currently, required to pay the fee for irrigation at the start of the season, which is the time when they need to invest in their business, rather than at the end when they have funds from the sale of their harvest. Although, there was some suggestion that the a priori fee, was in fact the ex post fee for the previous year. In addition, it is technically not possible to disconnect on the individual basis. Also, disconnections happen as a result of the poor organization of the payment arrangements. Decentralization issues 24. As provided in the new organic Law on LSG Units, abolishment of Rayon (district) offices will facilitate removing impediments to the development of effective local administrative units and increase decision-making power of local self- governments. However, the central government should ensure that transferring of responsibilities for rural infrastructure is aligned with adequate local capacity (as discussed in para 14). Formula-based equalization, potentially to be applied from 2007, will ensure more objective allocation of resources between the local self-government units. But on the other hand, the state should also ensure the adequate level of the equalization fund annually to make sufficient resources available for each unit so that adequate resources are available for maintenance needs. The level of equalization grants will greatly influence capacity of local self-governments to provide investments for rehabilitation and maintenance of rural infrastructure. The sustainability of interventions must be ensured. It is also important that integrated local budget reporting system is established in the Ministry of Finance and respective financial management capacity is established at the local government level to ensure effective planning, implementation and reporting system. 25. The new territorial-administrative setting of the country is expected to have positive impact on management and maintenance of rural infrastructure. As provided by the new Law on LSG Units, Georgia implemented the new administrative-territorial framework (effective November 2006). Two level administrative framework was applied consisting of central government and local self-government levels.6 Consolidation of former 1,100 small and resourceless local government units into 69 units7 may suggest increased professional capacity and know-how to be available in each administrative unit. Rural units 6See Annex B for detailed description of the institutional arrangements in rural infrastructure in Georgia. 7New 69 units consist of 59 former Rayons, 4 former Districts, and 4 special status cities - Tbilisi, Kutaisi, Batumi, Rustavi, and Poti. Local councils are elected in these 69 local self-governments which encompass 3,736 settlements (cities, towns and villages). 7 will benefit from such amalgamation if transparent and objective criteria are established for decision-making on prioritization of needed interventions and investments, as well as if adequate financing is ensured for management and maintenance of rural infrastructure. Box 1: Intergovernmental Relations in Transition Countries In many transition states, fragmented and weak self-governing authorities have been submitted to stringent controls by the state administration, through formal or informal means.(*) This fact has been recognized as an impediment to the development of effective local self-governing authorities and the countries have attempted to correct the distortions. In Slovakia, for instance, the former district offices, which were the main instrument of state control, are being abolished following the adoption of far reaching decentralization laws in 2001. While regional state administration remains, they no longer have the same influence and powers over local self-governing authorities. Also, the creation of elected regional authorities has further altered the balance between state administration and self-governing bodies. In Bulgaria, a reform of the country public administration has reduced the formal powers of state over decision- making of municipalities. Although serious issues remain in the system of intergovernmental finance, municipalities no longer work under strict controls of the state administration. (**) (*) "Innovations in Public Management," by Verheijen and Coombes, Edward Elgar publishers, 1998. (**)"Bulgaria, Issues in Intergovernmental Relations," World Bank 2004 Source: World Bank, PREM, Moldova Intergovernmental Fiscal Relationships, 2002 8 Box 2: Principles and Practice on Expenditure Assignment There is no fixed, established rule on how expenditures should be assigned among government levels. The extent and nature of decision-making power exercised by lower level authorities has varied considerably from country to country and may change from time to time, depending on technology and preferences. Especially in unitary countries (with centralized government), subnational governments are not constitutionally empowered to make independent decisions over a range specified public functions or services. However, general principles can be derived from best practices which indicate that policy decision on economic stabilization and income distribution are typically assigned to central government, while those related to allocative efficiency may be assigned to the local governments. Otherwise, when decisions on economic stabilization and income distribution are left to the local governments, wrong incentives and conflicts may arise, and policies maybe rendered unsustainable. Observance of the principle of subsidiarity--i.e., decisions taken at the lowest possible level of government consistent with allocative efficiency, including the internalization of the benefits within the geographic area and the provision of services at minimum cost--is fundamental to define the scope of local governments. More specifically, the national government in general retains responsibility for the provision of national public goods, international affairs, monetary policy, regulation, distribution and transfers to persons and businesses, fiscal policy coordination, regional equity, and preservation/promotion of internal common market. The intermediate level governments may be responsible for education, health, social insurance, inter-municipal issues, and oversight local governments. Local governments may take the responsibility for some local services, including the provision of utilities. Beyond areas of exclusive functions/responsibilities which are attributed to each government level, the law should clearly specify competences on shared functions. In this regard, central government should keep decisions on overall policies, setting standards, and auditing, while intermediate governments should be given competence on over-sighting implementation, and local governments should be more directly involved in providing/ delivering the services (and infrastructure). Assignment of respective responsibilities to regional vis-à-vis local governments should respond to country conditions and be dependent on considerations such as economy of scale (arising from the indivisibilities and specializations), economy of scope (arising from related activities), cost-benefit spillovers, proximity to beneficiaries, consumer preferences, and flexibility on expenditure priorities. In this regard, because of different local population size, distinct rural and urban characteristics, and diverse fiscal capacities, the assignment of responsibilities could be asymmetric, where larger local governments may end up having relatively larger responsibilities and higher degree of autonomy. Source: World Bank, PREM, Moldova Intergovernmental Fiscal Relationships, 2002 There is a need for greater clarity of ownership and responsibilities between levels of government 26. The study findings indicate that clear delineation of infrastructure ownership, and roles and responsibilities for funding, service delivery, operation/maintenance and technical oversight between central and local governments and infrastructure entities is critical for the sustainable development of rural infrastructure services. This delineation should take into account the type of the infrastructure and the preferences of the communities. 27. The ownership of the rural infrastructure should be delineated based on the type of infrastructure, with government ownership preferred in those cases where the infrastructure services have significant public good characteristics. These services may not be financially viable, reflecting the fact that the private sector is unlikely to consider the associated externalities. The government ownership should be further delineated between the central and local governments contingent on the extent the infrastructure serves national vs. local 9 needs. Thus, community ownership should be preferred for the infrastructure serving only local community needs while central government ownership would be appropriate in the case of common infrastructure serving several communities. 28. In general, the central government should not have rural infrastructure operation and management responsibilities. These responsibilities should be shared between communities and the private sector with communities assuming operation and management of infrastructure involving low technology and low costs, such as in the case of water and irrigation. Whilst more complex infrastructures, involving greater costs, technical abilities, such as with roads, or where there is a greater safety risks, such as with electricity and gas, the network provider should remain responsible. Communities should have the flexibility of outsourcing the Management, Operations and Maintenance (MOM) functions to the private sector in the future. However, there are concerns about the involvement at this moment of the private sector in MOM and excluding lower-order infrastructure from rehabilitation programs. Private sector water suppliers are for-profit companies and therefore they have to charge the real MOM costs for irrigation water as well as a profit. In contrast to private operators, not-for-profit farmer-owned associations have proven much more efficient and able to provide services at lower costs. This is why farmers should preferably have a major responsibility for the management of irrigation and drainage systems. There is a need for appropriate maintenance arrangements 29. The construction, rehabilitation and, to some extent, the maintenance of some of the rural infrastructure has already been out-sourced to the private sector. Despite some shortfalls, private owners and operators maintain, rehabilitate, and expand rural gas, telecom, and electricity infrastructure. 30. The communities should be familiarized with the low cost technologies that can be applied to those types of rural infrastructure remaining under community ownership. In the case of the road infrastructure, cost sharing arrangements between the Road Department of the Ministry of Economic Development (MED) and the communities could be established for the routine maintenance of the intra-community roads and roads connecting communities to the main network. 31. Most communities have adequate tools, equipment and human resources to carry out routine maintenance. With some additional guidance, these communities could contribute to cost effective maintenance of rural community roads. The study has found that the communities are willing to provide in-kind contribution to supplement the funding received from the state. There is a need for a champion of reform 32. The initiation and formulation of any CDD project is usually the responsibility of the community leader, who has to find financing, communicate with program managers, provide guidance to the community, organize work, communicate with the contractors and organize maintenance. The importance of championship carried out by individual(s) with incentives to push the infrastructure investment projects through all stages from financing to day-to day 10 maintenance of assets cannot be overstated. It is clear from past experience with rural investments in Georgia that a community with a strong leader, especially those with the support of the full community (which creates incentives for them to act), who play the role of a champion, are much more likely to succeed in project implementation than the average community. 33. However, recognition of this fact creates a policy dilemma; what to do in those communities where a community leader is absent, but which may be the most appropriate places for rural infrastructure investment, based on any assessment of cost effectiveness, or need. One option might be to use strong leadership as a sifting device in the early stages of project preparation, with `needy' communities without strong leadership, that are not selected for the first tranche of interventions within a program, to be provided with appropriate support to mobilize leadership, before any further investments in rural infrastructure can be made in those communities. The defined communities could then receive technical assistance to try and develop their capacities. 34. It is recommended that the Government support local leadership and participation in the decision process and, therefore, set aside some public money for community mobilization and technical assistance as part of the developed strategy. This is particularly important in places where inequality is large, or where the lack of leadership is an issue. HOW SHOULD RURAL INFRASTRUCTURE BE MANAGED? 35. Rural infrastructure is no different from any other infrastructure network, requiring professional management to ensure that the assets of the sector are maintained in a cost effective and efficient manner. Whilst each infrastructure network could be managed separately, and is for primary and secondary networks, there are strong grounds to try and seek synergy in investments, and economies of scale and scope, in network management in rural space. Where the network, or networks, has deteriorated significantly, the trade-offs between rehabilitation, upgrading and replacement across all the infrastructure types become even more difficult. 36. An approach of this type would require the development and use of a consolidated network planning tool. The development of such a tool would require information to be collected, and updated on a regular basis, on the location and condition of all rural infrastructure types. The required interventions to maintain the network, clear backlog maintenance, and possibly even extend the network would then need to be identified and costed. The final step would require the prioritization and planning of interventions, to ensure that the `finite' budget available is expended on those investments that provide the highest return, and that implementation is undertaken effectively. 37. This study has developed a Geographic Information Systems (GIS) database and a `Decision Matrix' methodology, based on a spreadsheet, (which together have been entitled the `Rural Infrastructure Assessment Toolkit' (RIAT), to assist the Government in managing, planning and prioritizing between different infrastructure interventions across sectors, across regions and differing rural communities. The following paragraphs describe their establishment and potential use in more detail. 11 The development of the GIS database 38. All the data that were collected in the study have been collated into a GIS database. This includes the infrastructure condition data for local roads, gas and electricity networks, irrigation and drainage, and telecommunications coverage. Digital maps on network location and photographs of structures have also been included, where available, along with other census and economic data, to create a powerful tool. The database has been constructed to cover all rural communities in Georgia. 39. The GIS database and associated software, as well as providing the base data for the Decision Matrix, can also be used both to indicate, in a visual format, the location and condition of the different infrastructure networks in Georgia, the location of earlier interventions and, hence, be employed for network management. 40. The GIS database could also, potentially, have a significant impact on the identification of linkages between the different sectors of rural infrastructure. By combining information on all infrastructure networks, GIS software can be used to identify areas were synergies in infrastructure development can be exploited as well as areas that are just missing "one small piece of the puzzle." The development of the Decision Matrix 41. This section describes the development of the pilot `Decision Matrix' methodology, which has been developed to assist the Government in prioritizing between different infrastructure interventions across sectors, across regions and differing rural communities (the Decision Matrix methodology is outlined in Appendix 1 of this report). The Decision Matrix is a `spreadsheet based' multi-sectoral assessment tool, using data from the GIS database, and the community preferences of 100 communities to allow the prioritization of interventions in that sub-set of communities8. 42. The broad approach is a form of stylized multi-criteria analysis, where all potential investments are assessed, and provided with a normalized score, against a number of pre- defined criteria. These scores are then aggregated to facilitate the ranking of the investments, and provide a measure of the relative `value' of each against all other prospective investments. An explanation of each criterion is presented in Appendix 1. 43. The criteria included in the Decision Matrix were explicitly selected to reflect both the long and short term strategic objectives of the government, such as poverty alleviation, economic growth and the cost effectiveness in the use of scarce public resources. In addition, the Decision Matrix also incorporates the preferences of communities thus, to some extent, also addressing affordability and sustainability issues. 8The Decision Matrix currently only covers the 100 communities where community preferences were gathered during the course of this study. 12 44. The number of people registered as unemployed is used as a proxy for poverty due to the paucity of poverty data at the LSG or rayon level9. The formula used to calculate cost- effectiveness is bulleted below, and was selected to try and alleviate potential distortions where the size of the population was small, but the absolute number of poor households was large, relative to comparative communities: number of people /cost of intervention + number of poor/cost of intervention 45. The intervention cost for all communities and types of infrastructure has been estimated in the course of this study (for more details see Chapter on Findings of the Study, section on "Estimated Cost to Rehabilitate the Rural Infrastructure"). It is important to mention that these costs are based on the current technologies, standards, and unit prices prevailing in Georgia. Also, the interventions are based on the principles of provision of MLS, although not explicitly defined by the Government. In some cases, significant cost reductions could be realized as a result of introducing more cost-effective technologies, streamlining excessive standards and formulating the concept of more realistic and affordable minimum level of service. 46. Finally, the potential for growth in each rayon has been estimated by a number of national experts, using the Delphi method. The intention of the exercise was to use their expertise to make sure that those areas that are considered to have the highest growth potential, are represented additionally in the Decision Matrix, to provide the Government with the option to focus interventions in areas of highest economic potential. How would the Decision Matrix be used? 47. There are a number of ways to use the Decision Matrix; firstly, the Government may decide to retain the broad approach, and focus the finite budget on those investments that offer the highest cumulative return. Alternatively, the Government may wish to focus on poverty alleviation, in which case the relevant button in the Decision Matrix is pressed, focusing investments in those communities which have the highest levels of `poverty', as measured by the defined proxy. The differences in the output for the different possibilities are reported in the following section, whereas an example of the output from the Decision Matrix in tabular form is presented in Appendix 2 of this report. WHAT WOULD BE THE NEXT STEPS IN THE ESTABLISHMENT OF THE RIAT? 48. The World Bank currently has a program that seeks to establish GIS based systems in client countries, involving the use of capacity building devices to Government agencies and other institutions to operatize the systems. A recent review of the use of asset management systems in the highway sector10 found that there are three key elements to the sustainable introduction of such a system: technical aspects, the system itself and whether it is appropriate for the context; processes, establishing the processes that will use the system; and 9This measure is acknowledged to have its flaws, but in the absence of a more robust proxy, was selected for use in this study. 10See McPherson, K. and Bennett, C. (2006) for more information. 13 finally the people, ensuring that training is provided to suitable individuals within the stakeholder institutions. For technical requirements see Annex A.11 THE OUTPUT FROM THE DECISION MATRIX 49. The following figure reveals the composition of the highest scoring fifty investments across all sectors and all communities. Drinking water is the most dominant infrastructure intervention, accounting for 44 percent of the total. This is followed by natural gas and telecommunications, with 18 percent each, and electric power with 12 percent. Roads and Transport are the least dominant infrastructure interventions making up only eight percent of the 50 highest scoring interventions, reflecting the high proportion of secondary roads in the road network connecting the communities, with ownership and management clearly delineated to the Roads Department of the Ministry of Economic Development. Figure 1 : Sectoral Composition of the Highest Scoring Interventions Telecommunications 18% Roads/transport Drinking water 8% 44% Natural Gas 18% Electric Power 12% 50. If the primary emphasis is placed on poverty reduction, then electric power ranks the highest, representing 28 percent of the sector priorities in the top fifty ranked interventions. This is followed by drinking water, with 20 percent, and roads and natural gas, with 16 percent each. 11Annexes are provided in Volume II of the report, available on CD ROM. 14 Figure 2 : Sectoral Composition of the Interventions with Highest Emphasis on Poverty Reduction Telecommunications 14% Drinking water 20% Roads/transport Irrigation 16% 6% Natural Gas Electric Power 16% 28% 51. If the primary emphasis is placed on economic development, electric power ranks the highest, representing 28 percent of the sector priorities in the top fifty ranked interventions. This is followed by drinking water, with 22 percent, natural gas, with 20 percent, and irrigation with 12 percent respectively. Figure 3 : Sectoral Composition of the Interventions with Highest Emphasis on Economic Development Telecommunications 8% Drinking water 22% Roads/transport 10% Natural Gas Irrigation 20% 12% Electric Power 28% 15 The policy implications of the output from the Decision Matrix 52. The implication of the output from the Decision Matrix is that if there is a desire on the part of the Government to focus exclusively on economic growth or poverty reduction, then the primary emphasis in any action plan, or component therein, for rural infrastructure investments would be on the provision of electric power and drinking water.12 The priorities areas, however, may vary for different locations. HOW MUCH WOULD THE INVESTMENT PLAN COST AND HOW WOULD IT BE FINANCED? 53. There is a clear need for a robust financing plan indicating needs to donors. It is recommended that the Government define a financing plan for the sector consistent with the Medium Term Expenditure Framework (MTEF). The resulting plan needs to be realistic, clearly indicating the portion to be provided by the Government, the proportion, and type, of contribution required from the communities, together with firm possibilities of private sector contribution, where feasible. It would also need to indicate the portion of the funding sought from donors, to allow the latter to contribute, within the framework defined by the Government. How much would the investments cost? 54. The estimate of this study is that approximately US$1000 million would be required to rehabilitate the rural infrastructure networks in the whole of rural Georgia. This range includes about US$212 million for the rehabilitation of "lifeline" roads, including considerable lengths of secondary roads, US$439 million for the supply of gas to the majority of rural communities, about US$87.3 million for electricity, about US$46 million for telephony, depending on assumptions made about the sophistication of the required telecommunication, about US$47 million for potable water, and US$165 million for irrigation and drainage (The basis for the estimates is provided in Chapter on Findings of the Study, section on "Estimated Cost to Rehabilitate the Rural Infrastructure"). 55. It should be stressed that these figures are intended to be indicative, rather than definitive, cost estimates, as they are based, largely, on current prices, which predicate current standards, conventional technologies, the principles of provision of MLS implicitly assumed by the government, and proxy information where unit costs were unavailable. The actual costing would very much depend on: (i) the policy choices, or trade-offs, that are made in respect of provision of MLS; (ii) the technologies that are used for the rehabilitation and development of the infrastructure; (iii) the level or quality of the infrastructure services to be provided; and (iv) the viability of the individual investments. 12The priority areas, however, may vary if the focus is poverty reduction or economic development. In case of Georgia the priorities are the same if either of these two criteria is emphasized, however, they are different for different locations. 16 Can the Action Plan be financed from the local budget? 56. The rural communities are heavily dependent on transfers from the central budget, which typically comprise over fifty percent of local budget revenue. Other sources of revenue for the community budget are property and land taxes (assigned to and collected by the local governments), certain types of state duties, local duties and other non-tax revenues (rent, fines and penalties, sale of community owned property, etc). 57. There are real limits as to what can be raised from local taxes, implying that if a program of rural infrastructure investments was to be undertaken in whatever form, there would need to be a significant transfer of capital expenditure to the local communities, together with increased provision of recurrent expenditure to ensure the sustainability of any investments. Nonetheless, the evidence from the WTP study suggests that the community contribution could be significant in some cases, and would need careful assessment at the appraisal stage. In addition, there is a need to ensure that both current, and future, funding is used efficiently and effectively. What can the private sector do? 58. While it is unrealistic to expect that the private sector will finance the provision of MLS for all rural residents (and as discussed earlier in the text in that regard significant policy changes have to occur) the private sector can play an important role in financing the provision of rural infrastructure services. 59. The private sector, possibly with support from the public sector, can design or extend the existing credit and/or micro-financing schemes for communities, as well as individual community members, to improve infrastructure services, e.g. gas, water, communications. These schemes, in particular, can finance connections of individual households to gas, telecom and water network. Approximately 50 percent of households in the communities that have gas infrastructure do not use gas. This is because they can not afford to pay the large fee for the connection to the gas network. A similar situation can be found in the telecom and water sectors. Also the private sector will be expected to contribute, through a provision of MLS financing mechanism, in the manner described earlier, to the provision of MLS. 60. Finally, private utilities are often willing to finance the extension of the infrastructure to a particular community, if there is a parallel commitment from the community, or the government, on the intra-community network and individual household connections. This provides an opportunity for public-private partnership in the provision of some of the services. The evidence from the WTP surveys would suggest, based on the sample, the potential cost savings from the provision of gas and the reported WTP of households is of a scale sufficient to ensure a role for the private sector in service provision. This assertion is supported by the reported size of some of the community contributions under the CDD projects. What can the donors contribute? 61. Some information about the scale and scope of potential donor support over the next five years is available. Over the 2006-2009 period, the World Bank expects to disburse 17 approximately US$50 million to support rural development, US$78 million to support the energy sector, and a further US$84 million to develop transport infrastructure in Georgia. The World Bank Secondary and Local Roads Project (US$27 million), currently under implementation, is helping to address bottlenecks to rural development through rehabilitation of approximately 500-750 km of paved `secondary and local roads, including 150 to 250 km of local roads. 62. Other relevant projects currently under implementation, supported by the World Bank, include the Irrigation Drainage Community Development Project ($27 million), and the Georgian Social Investment Fund II ($15 million). The Irrigation Drainage Community Development Project is assisting to raising the productivity of land for high-value crops in targeted rural areas through rehabilitation of main irrigation and drainage infrastructure. It will also help restore irrigation infrastructure and river embankments damaged in the recent floods and strengthen longer-term flood protection measures. 63. The GSIF I (which has now closed) and II (ongoing) target mobilizing public and private funds for community-driven basic infrastructure projects in poor rural communities and strengthen community level capacity to prepare and implement investment projects. Kreditanstalt Für Wiederaufbau (KfW) is also funding GSIF for implementation of rural infrastructure rehabilitation projects. Over the last decades United States Agency for International Development (USAID) financed community projects have also been targeting small community grants programs for rehabilitation of small scale rural infrastructure (road, water, etc) in targeted areas. 64. The Millennium Challenge Corporation (MCC) represents another major source of investment into Georgia's infrastructure as Georgia was awarded a grant of US$295 million in 2005. The Program was officially launched in April, 2006. From this grant amount $49.5 million is to be spent on the energy sector, including repairs to the gas trunk pipeline, about $100 million on road improvement work, $60 million on regional infrastructure development, $32.5 million on financing development investment funds and $15 million on developing the agro-industrial complex. This program is being implemented by the Millennium Challenge Georgia (MCG) and, given the emphasis on rural areas, is expected to have a significant impact on rural infrastructure in Georgia. 18 KEY RECOMMENDATIONS 65. Suggested recommendations are summarized in the following matrix: Broad Area Actions Key Issues to Address x Define the investment needs and priorities across sectors, regions, and communities Development of a coherent Rural x Identify critical parts of the network to maintain Infrastructure Strategy (RIS) and Action Plan x Explore public-private partnership opportunities x Clarify the issues related to local ownership, maintenance, regulation and standards of infrastructure assets x Ensure alignment of government and donor resources, policies and Rural Increased coordination between actions with the overall Government strategy Infrastructure donors under Government x Designate one government institution to coordinate donor-funded Development ownership activities in rural infrastructure Strategy x Develop robust financing plan indicating needs to donors x Increase Access to infrastructure services in poor undeserved areas Development of a mechanism to x Define concept of MLS in the legislation for all sectors finance Minimum Level of Service (MLS) x Provide direct subsidies from the state budget or establish shared financing mechanism for MLS x Outsource management and operations in the provision of public Exploring public-private services partnerships x Design and implement minimum level of service (MLS) provision schemes x Collect and update on a regular basis the information on the location and condition of all rural infrastructure Development of consolidated x Identified the required interventions to maintain the network network planning tool x Prioritize and plan interventions to ensure highest return on the limited budget Improved Management x Assess the technical aspects of the system to identify whether it is appropriate for the context of Rural Infrastructure x Provide technical assistance to facilitate the establishment of the Establishment of Rural system to identify required changes for the installment and use of Infrastructure Assessment Tool the system (RIAT) in RDMED x Train the operators; ensure that training is provided to suitable individuals within the stakeholder institutions. x Develop GIS Database and the Decision Matrix x Create incentives to pay by improving quality of services Improvement of billing and metering systems x Improve efficiency of management and organization of the payment arrangements x Define clearly roles and responsibilities for funding, service Delineation of infrastructure delivery, operation/maintenance and technical oversight between Sustainability ownership and responsibilities central and local governments and infrastructure entities of Rural between levels of government x Consider community preferences and private sector Infrastructure Investments x Delineate ownership based on type of rural infrastructure Establishment of appropriate x Outsource to private sector maintenance arrangements x Encourage community ownership of low cost technologies Support of local leadership and x Allocate public money for community mobilization and technical participation in decision making assistance, especially for places with lack of leadership process 19 20 THE FINDINGS OF THE STUDY WHAT IS THE COVERAGE AND CONDITION OF RURAL INFRASTRUCTURE? The Electricity Sector Access to and availability of electricity 66. The survey of local communities, undertaken as an input to this study, found that 97.3 percent of the rural population has access13 to electricity, however, availability14 is low. Across the country as whole, the average household in a rural area is supplied with electricity for between six and eleven hours per day. However, there is considerable variation in supply, with electricity available for more than six hours a day for only 30 percent of rural households, from three to five hours a day for 36 percent of rural households, and 32 percent of rural households having access for less than two hours a day (see Figure 4). 13Defined as an existing physical connection, whether operational or not. 14Defined as: (a) an operational connection; or (b) service is provided by the supplier. 21 any 22 of ust sta legal the on nte judgm nya roup, G nk Ba World the of part the on ply, im not do pa m this on howns (2005) ation rayon rmo by inf boundaries. day other such per any of and supply ty nationsi acceptance or ent electrici denom,s of loroc endorsem data Hours any : 4 or Study boundaries, Figure The territory, Source: 67. The average monthly household consumption of electricity was revealed to be 250 kWh, but with significant regional variation. Kakheti region in the west had the lowest average monthly consumption of 75 kWh, whilst Shida Kartli region had the highest with an average monthly household consumption of 500 kWh, as revealed in Table 1. Table 1 : Access, availability and consumption of electricity within households, by region Access to electricity, Availability of electricity, % of Household monthly % of households households with access consumption of 6 or more 3-5 hours 2 hours a electricity, kWh hours a day a day day or less Adjara 93.9 NA NA NA NA Imereti 97.6 40.0 37.9 22.1 333 Guria 72.8 48.6 29.9 21.5 250 Samegrelo 89.1 0 56.4 43.6 333 Shida Kartli 96.0 65.7 34.3 0 500 Racha-Lechkhumi 98.1 100 0 0 250 Mtskheta ­Tianeti 96.1 39.3 38.8 21.8 333 Kakheti 96.1 14.8 59.4 25.8 75 Samtskhe-Javakheti 96.4 44.9 55.0 0 166 Kvemo Kartli 97.2 36.3 0 69.2 250 Rural Georgia 93.7 29.5 36.1 34.4 250 Source: Study data 68. The relatively high average consumption level and the high variance can be explained by supply rationing, and poor billing and revenue collection. Electricity is not metered at the household level in rural areas, and households are billed a fixed monthly rate of about GEL 5 (US$2.5) each, providing little incentive to individual households to reduce consumption. In the 30 days prior to the interview, only 19 percent of communities had uninterrupted supply, 61 percent were supplied according to a previously determined schedule, while 20 percent had irregular supply interruptions. 69. Respondents were also asked if the quality of the power supply has deteriorated since the pre-transition year of 1990: in 73 percent of surveyed communities electric power supply was considered to have worsened. It was considered unchanged in 12 percent of communities, whilst 15 percent considered that supply had improved. The sources of electricity supply 70. The data reveals that electricity is supplied by a public company in 94 percent of the communities, whilst the remaining six percent receive it from a private company. In the latter cases, the power is produced by a small local power station or by generator, either at the household or at the community level. However the unreliability and/or rationed power supply from the public company means that 23 percent of those communities that have 23 access to mains supply, also have a secondary source of supply, either in the form of a private company, or a generator. 71. These latter options are clearly inefficient and expensive methods of ensuring supply, either provided by the community or by a private company. The provision of capacity requires a substantial upfront investment, and revenues are low, as the tariffs are required to be set at the same level as the public sector supplier (GEL 5 per month), which receives significant subsidy. The period of supply of electricity to communities, by source of supply, is provided in Table 2. The number of hours a day when electricity is supplied by the public company is defined by rationing and supply interruptions, the remaining two sources are utilized mainly as substitutes. Table 2 : Supply of electricity to communities by electricity provider (% of communities) Average Period of Supply Source of Supply 6 hours a day or less 7 - 12 hours a day 13 - 24 hours a day public company 39% 26% 35% private company 26% 37% 37% own generator 94% 3% 3% Source: Study data The condition of the electricity network 72. Respondents in only 0.1 percent of communities (1/100) responded that the electricity network was in good condition, while respondents in 33 percent of communities think that the network requires major repairs, and 67 percent thinks there is a need for immediate major repairs, with wide variation by region around the average (see Table 3). 24 Table 3 : Condition of electricity network (% of communities) Region Good Requires Major Requires Immediate Major Repair Repair Adjara NA NA NA Imereti 0.1 49.3 50.0 Guria 0 58.0 42.0 Samegrelo 0 14.2 85.8 Shida Kartli 0 81.4 18.6 Racha-Lechkhumi 0 72.3 27.7 Mtskheta ­Tianeti 0 3.8 96.2 Kakheti 0 99.0 1.0 Samtskhe-Javakheti 0 24.7 75.3 Kvemo Kartli 0 30.8 69.2 Rural Georgia 0.1 33.1 66.8 Source: Study data The Gas Sector Access to and availability of mains gas supply 73. The data for the access and availability of the supply of mains gas is provided in Table 4. On average, across all the regions, just over 20 percent of households have access to the gas supply, whilst only just over three percent are actually being supplied with gas in their homes. Three regions, Adjara, Guria and Samegrelo have no access to mains gas, whilst in Imereti only 3.3 percent of communities and households have gas supply. On average, 81 percent of LSGs are within a distance of 50 km or less from the nearest main gas pipeline, but Samegrelo is something of an outlier as only just over 42 percent of communities lie within the same threshold. 25 Table 4 : Access to Gas Supply System Region Have access % HH Distance to the nearest Gas Supply Pipeline to Gas being supply supplied <50km 50-100km >100km % of HH Adjara 0 0 64.5 35.5 0 Imereti 3.3 3.3 96.2 3.1 0.6 Guria 0 0 100 0 0 Samegrelo 0 0 42.2 43.7 14.1 Shida Kartli 70.9 1.2 91.4 0 0 Racha-Lechkhumi 0 0 na na na Mtskheta ­Tianeti 46.7 10.8 83.0 5.7 0 Kakheti 37.4 1.8 99.2 0.8 0 Samtskhe-Javakheti 11.8 8.6 59.0 41.0 0 Kvemo Kartli 17.4 31.0 84.6 15.4 0 Average rural Georgia 20.65 3.29 81.0 16.6 2.4 Source: Study data 74. The following figure provides a visual indication of the distribution of those Sakrebulos that currently have natural gas supply. 26 any 27 of ust sta legal the on nte judgm nya roup, G nk Ba World the of part the on ply, im not do pa m this on howns ation (2005) rmo inf boundaries. Supply other such any of Gas and Natural nationsi acceptance or with ent denom,s loroc endorsem data Sakrebulos any : 5 or Study boundaries, Figure The territory, Source: The condition of the gas distribution network 75. The condition of the main gas systems in rural Georgia is very poor, with an average of 86 percent of all communities having a mains gas system that is in very poor condition. Again, there is variability, with four regions having 100 percent of their respective systems in very poor condition. By contrast, on average only 19 percent of the mains gas system is in good condition, driven mainly by the condition in Racha-Lechkhumi and Mtskheta-Tianeti. The average gas consumption of those households being supplied with gas is 19 kbm. Table 5 : Condition of Gas Supply System by community and region Region Average % in Good % in Poor % in Very Poor consumption per Condition Condition Condition Household (supplied Kbm) Adjara 0 0 0 100.0 Imereti 62.9 4.4 3.1 92.5 Guria 0 0 0 100.0 Samegrelo 0 0 0 100.0 Shida Kartli 3.99 1.4 24.3 74.3 Racha-Lechkhumi 0 0 0 100.0 Mtskheta ­Tianeti 24.75 3.8 13.2 83.0 Kakheti 22.9 1.8 53.2 46.0 Samtskhe-Javakheti 9.8 2.5 1.2 96.3 Kvemo Kartli 1.2 5.8 16.3 77.9 Average rural Georgia 19.0 2.0 12.3 85.7 Source: Study data The Drinking Water Sector 76. The survey data reveals that 96 percent of rural households have access to drinking water, while only 27 percent have access to centrally supplied water. In addition, supply even where provided is not perfect, with frequent interruptions and quality variations. This reflects the poor condition of the assets in the sector, and the lack of substantive reform. The survey data reveal that only 64 percent of those who receive centrally supplied water pay anything for it, with a mean monthly charge of GEL 1 per month. Access to water from different sources 77. In schools and local markets, drinking water is accessible in 61 percent of Communities. Most of Communities (62 percent) are supplied with central water, and in these Communities households have access to it either through a tap in the house or in the yard or through street stand pipes. The rest use well water (33 percent of communities) and springs (4.5 percent of communities) (see Table 6). 28 Table 6 : Access to water and main source of water by region (% of households, % of communities) % Households with Access Main source of water Drinking tap water drinking drinking Central (house tap, Well, % Spring, % Water water in water in yard tap, street commun communiti schools market stand pipe), % ities es communities Adjara 96.7 53.6 95.1 95.1 93.4 6.6 0 Imereti 81.1 27.6 68.6 68.6 68.6 30.9 0.6 Guria 100.0 10.4 54.2 54.2 61.0 0 39.0 Samegrelo 100.0 13.4 41.2 41.2 41.9 10.8 47.3 Shida Kartli 100.0 17.5 60.0 60.0 60.0 40.0 0 Racha-Lechkhumi 100.0 39.2 55.4 55.4 55.4 10.8 33.8 Mtskheta ­Tianeti 100.0 24.9 38.5 38.5 43.4 9.4 47.2 Kakheti 100.0 21.9 47.2 47.2 64.2 35.8 0 Samtskhe- 100.0 20.9 70.6 70.6 70.6 17.6 11.8 Javakheti Kvemo Kartli 100.0 53.5 78.5 78.5 73.8 0 26.2 Rural Georgia 96.1 27.1 60.5 60.5 62.4 33.1 4.5 *Source: Survey data. 78. The community survey, conducted as part of this study, reveals that only 29 percent of the sampled Georgian communities have access to centrally supplied water, with only 25 percent considering it as their main source of water. The remaining communities use wells, stand pipes or natural sources (springs, rivers, lakes) as their main sources (respectively 36 percent, 10 percent and 28 percent of surveyed communities). The following figure indicates the distribution of schools without water supply: 29 any 30 of ust sta legal the on nte judgm nya roup, G nk Ba World the of part the on ply, im not do pa m this on howns (2005) ation rmo inf boundaries. Sakrebulo other such by any of and Supply nationsi acceptance or Water ent with denom,s loroc endorsem data Schools any : 6 or Study boundaries, Figure The territory, Source: Availability of water and water quality 79. However, both scheduled and irregular interruptions in central water supply during the day means that even those communities that have access to central water, and those that consider it as their main water source, are required to use alternative sources of water: wells, stand pipes and springs/rivers. Accordingly, four percent of surveyed communities have central water access, but don't consider it their main water source. A further 25 percent of surveyed communities that use central water as their main water source, more than half of these communities regularly use other sources of water, mostly natural springs and wells, as substitutes. 80. Data collected for this study reveals that 69 percent of respondents in communities with access to a mains water supply reported an uninterrupted supply. A further 23 percent reported supply only during scheduled hours, typically for eight hours only, with the remaining eight percent reporting seasonal supply only. 82 percent of respondents receiving mains water reported that the quality of water from that source was good. The corresponding proportions for water from wells, stand pipes and springs/lakes was 75 percent, 34 percent and 25 percent respectively. Condition of water supply 81. Figure 7 indicates the water and sewage system in Georgia, and the proportion that is considered to need immediate repair, based on the qualitative responses of respondents in the WTP survey (N=1000). 31 any 32 of ust sta legal the on nte judgm nya roup, G nk Ba World the of part the on ply, im not do pa m this on howns (2005) ation rmo inf Condition boundaries. other such any of and frastructure In nationsi acceptance or Sewage ent denom,s and loroc endorsem data Water any : 7 or Study boundaries, Figure The territory, Source: The Telecommunications Sector Access to telecommunication services 82. The survey of rural communities, conducted as part of this study, revealed that only 26 percent of surveyed communities have working telephone lines. In communities with working telephone lines, about 20 percent of households have a landline telephone service at home. The situation with cellular telephones is different: while almost all communities (94 percent) have access to the cellular network, only a very small percentage of households (seven percent) own cellular phones and thus are able to use the network. In addition, access to public telephone is also limited: only 42 percent of communities have public phones. The condition of landline network 83. The condition of the telecommunication infrastructure in rural Georgia cannot be considered good. The earlier survey conducted for the Rural Infrastructure Study revealed that only three percent of communities have telephone infrastructure that is in good condition, in 25 percent of communities it requires rehabilitation and repair, and whilst in 73 percent it requires immediate repairs with considerable investment. Table 7 : Condition of landline network (% of communities) Regions Condition of landline network Good Need Major Repair Needs Urgent Major Repair Adjara 0 73.8 26.2 Imereti 0 42.9 57.1 Guria 0 6.8 93.2 Samegrelo 0 4.1 95.9 Shida Kartli 37.1 0 62.9 Racha-Lechkhumi 0 3.1 96.9 Mtskheta ­Tianeti 0 58.5 41.5 Kakheti 0 9.4 90.6 Samtskhe-Javakheti 0 18.6 81.4 Kvemo Kartli 0 9.2 90.8 Rural Georgia 2.9 24.7 72.5 Source: RCIS 2001 84. Figure 8 reveals that on average 82.5 percent of LSGs and 76.8 percent of households have some access to the telecommunication, however telephone density measured as telephones per 100 households appears to be low and on average across the country only 13 households per every hundred have telephone lines installed in their homes. This indicator is different across the regions. The Racha-Lechkhumi region is characterized with the lowest 33 rural telephone density (0.01/100 HH), while the highest is observed in Kvemo Kartli region (21.0/100 HH). 85. Landline telephone connection is more common in the district centers and LSGs located nearby the district capital, where the phone stations are located, however the latter is also available in other LSGs, although their number is limited and varies from 3.1 percent of LSGs in Racha-Lechkhumi region to 64 percent of LSGs in Adjara region. The status of telecommunication infrastructure across rural Georgia is poor. Only 2.9 percent of total LSGs in the rural areas have telephone infrastructure in a relatively good condition, while 24.7 percent require major repair and investments, and 72.5 percent immediate investment in the system. By contrast, on average, 87 percent of rural areas have cellular coverage, as shown in the next figure, although, the extent of use by rural residents is unclear given cost and incomes in these areas. 34 any 35 of ust sta legal the on nte judgm nya roup, G nk Ba World the of part the on ply, im not do pa m this on howns ation rmo inf boundaries. (2005) other such any of and Coverage nationsi acceptance or ent Telephone denom,s loroc endorsem data Landline any : 8 or Study boundaries, Figure The territory, Source: any 36 of ust sta legal the on nte judgm nya roup, G nk Ba World the of part the on ply, im not do pa m this on howns ation rmo inf boundaries. (2005) other such any of and Coverage nationsi acceptance or ent Telephone denom,s loroc endorsem data Cellular any : 9 or Study boundaries, Figure The territory, Source: The Rural Transport System Length and accessibility of local roads 86. Georgia has a relatively small road network consisting of roughly 23,550 kilometers, of which the primary and secondary network comprises some 19 percent (4500 kilometers); the remaining 19,050 kilometers are secondary and local roads. Table 8 presents main sector indicators by region. Table 8 : Comparative Road Indicators by Region Regions Population Length Main and Rural Roads Density Road Secondary Network Roads 100 (km) (km) Length Density Density Persons per (km) (m/sq.km) (m/person) sq km Adjara 1.06 1,136 176 636 3.2 3.03 Imereti 0.83 2,975 564 2,411 39 21 Guria 0.90 919 129 790 6.7 7.5 Samegrelo 0.75 2,699 500 2,199 5.9 7.8 Shida Kartli 0.86 2,348 361 1,987 5.7 10.6 Racha-Lechkhumi 0.22 1,065 283 782 4.1 18.9 Mtskheta ­Tianeti 0.26 1,977 407 1,570 5.1 17.1 Kakheti 0.66 3,053 510 2,543 6.0 7.9 Samtskhe-Javakheti 0.34 1,595 377 1,218 3.7 8.0 Kvemo Kartli 0.11 2,697 534 2,163 7.6 7.9 Source: Rural Infrastructure Study (2004) 87. The study included a visual condition survey, and Global Positioning System (GPS) mapping, of the defined `lifeline' road infrastructure. The latter was defined as that proportion of the network that represented the shortest distance between any one community and the nearest road in the next highest category of road. So for any one community, it would be the shortest length of tertiary road to the nearest section of secondary road. All these lengths, which were identified in a GIS based system, were aggregated and amounted to some 5755 km of road, or 30 percent of the entire local road network. 37 any 38 of ust sta legal the on nte judgm nya roup, G nk Ba World the of part the on ply, im not do pa m this on howns (2005) ation rmo inf Georgia boundaries. in other such any of and Condition and nationsi acceptance or ent denom,s Network loroc endorsem data Road : any 10 or Study boundaries, Figure The territory, Source: 88. The survey data also reveals that the rural road network within the communities is generally in a poor, or very poor, condition, as shown in Figure 10 (a comprehensive list of the condition of all roads, by rayon, is provided in Annex H). In addition, a majority of roads are unpaved and have received little if any maintenance for some years: only in five percent of communities roads were repaired within the last five years, while in 41 percent of communities they were last repaired 15 or more years ago. Access to transportation services 89. The survey of rural communities confirmed the finding of the earlier Rural Communities Infrastructure Survey (RCIS) that transport remains a serious problem for many Georgian communities: respondents in 31 percent of surveyed communities reported that public transportation to the nearest market place was not available. Accordingly, they were reliant on lifts to get to a departing point for public transport (in the form of minibuses generally), or direct to the market. 90. The survey also revealed that in 74 percent of surveyed communities, there was no transport to the nearest health facility, and 50 percent of communities had no transportation access to the regional center. The data also shows that whilst public transportation is available in some form in many, but not all, communities, the proportion of communities with public transport to the nearest health facility was 24 percent, whilst the comparable figure to the nearest rayon capital is 36 percent. The Irrigation Sector The availability of irrigation 91. Irrigation is a necessity in Eastern Georgia, but needed less in Western Georgia, where climatic differences, notably the presence of convection rainfall, means that drainage is more important. According to the survey, 52 percent of Georgian communities require irrigation for the whole arable land area, whilst 34 percent of communities do not need irrigation at all, with the remaining 14 percent requiring irrigation for a part of their arable land. In 2/3 of communities in the former category, i.e. those requiring irrigation of 100 percent of their arable land, only 50 percent or less of the land requiring irrigation actually benefits from a functioning irrigation system, as revealed in Table 9. Table 9 : Percentage of irrigable land receiving irrigation in communities (units: % of surveyed communities) Percentage of irrigated land Communities in each group 25% or less 37% 26-50% 30% 51-75% 11% 76-100% 22% Total 100% Source: Study data 39 The condition of irrigation systems 92. The data collected during the course of the survey reveals that out of communities that have irrigation systems, only 15 percent report that their systems function effectively all the time, with another 19 percent reporting that their irrigation systems are working, although with interruptions. The remainder report problems with irrigation systems: they are in bad condition or don't work at all. Table 10 presents a comparison of the condition of the respective irrigation systems in 2001 with 1990 to reveal the extent of the deterioration. However, it should be noted that the comparison reveals nothing about the desirability, or the viability, of returning the whole network to operational condition. Table 10 : Condition of irrigation systems (units: % of surveyed communities) Irrigation system... Communities in each group Year 2002 (current) Year 1990 (retrospective) Operates uninterruptedly 15% 55% Operates with interruptions 23% 29% Is in bad condition 51% 12% Does not operate 11% 4% Total 100% 100% Source: Study data WHAT ARE THE COMMUNITY PREFERENCES IN RURAL INFRASTRUCTURE? 93. The qualitative survey of rural communities (N=100) was undertaken to ascertain community preferences in terms of improvements to their rural infrastructure, and to gain an insight into how those preferences changed, reflecting variation in the following key parameters: (a) territorial-administrative division; (b) distance from the highway; (c) community altitude above the sea; (d) community size (number of households); (e) state of the infrastructure in the community; and (f) type of infrastructure ownership. Preferences were solicited in focus groups and community meetings in the sampled communities, and an example of each such event is provided in Figure 11. Figure 11 : Example of Community Meeting and Focus Group conducted during the Study 40 94. For small communities, the number of participants was approximately 30 and for large communities 50 to 60. Community interview participants were chosen according to the following principles. The groups included women and men; people of young, middle and old age; and people from different income groups. The idea was to have sufficient representation in the meetings from all named groups. The detailed approach employed to select the sample is provided in Annex F. 95. The selection process was random and based on registration books. First, the step of selection was determined: to select 30 participants, the number of households in the registration book was divided by 30 and thus the step was determined. For example, there are 600 households in the registration book. The step will be 20, and every 20th household would be selected from the registration book. While inviting to the interview the representatives of those households, the study group took into consideration the gender and age of the people, in order to ensure the representation of those groups. The presence of the village authorities at the meeting was discouraged to avoid their influence on opinions expressed. These principles were followed by the study team, however, there were some problems: x In Racha-Lechkhumi region, some Sakrebulos are fully inhabited only in summer. As the study field work was conducted in the winter, only men doing agricultural work were present in these Sakrebulos. x In Sakrebulos of Kvemo Kartli mostly inhabited by Azerbaijani, women's attendance was limited, as women are less active socially. 96. Respondents were asked to indicate which type of rural infrastructure was, in their opinion, the priority for rehabilitation. Table 11 reveals that electricity supply, gas supply and roads/transport were selected as the main priorities for the majority of communities in each of three regions. As an example, improved electricity was seen as a priority in just over 42 percent of surveyed communities in Guria, 34 percent in Samegrelo, and 23 percent in Shida-Kartli. Table 11 : Proportion of communities expressed preference for infrastructure rehabilitation, by type Kvemo Samtskhe- Shida- Guria Imereti Kakheti Mtianeti Kartli Racha Samegrelo Javakheti Kartli Drinking Water 12.0% 4.8% 28.5% 13.8% 15.0% 21.9% 4.3% 12.3% 15.9% Irrigation 0.4% 2.8% 4.8% 9.3% 13.1% 0.7% 0.2% 21.3% 19.1% Telecommunications 2.6% 13.9% 8.4% 15.9% 5.7% 17.3% 5.4% 9.6% 5.4% Natural Gas 26.4% 35.5% 18.9% 27.5% 23.1% 3.0% 21.4% 18.4% 23.0% Roads/Transport 16.5% 18.9% 11.7% 14.9% 22.8% 31.6% 34.4% 22.5% 12.7% Electric Power 42.2% 24.1% 27.7% 18.6% 20.4% 25.6% 34.4% 15.9% 23.9% Source: Study data 41 97. By contrast, irrigation was not reported to be the highest priority in any region, but it was important in some communities. In Shida-Kartli and Samtskhe-Javakheti, irrigation was the third and second most important priority at the regional level, but about 20 percent of communities indicated that it was their priority, in each case. In humid regions such as Guria, Racha and Samegrelo, very few people voted for irrigation. Drinking water was also not given the highest priority, except in Kakheti and Racha regions, although it was acknowledged as important by all. The relative abundance of natural sources of water in Georgia is probably a factor behind the low importance allocated to water supply. Improved telecommunications was also not seen as a priority in any region. 98. There was also a significant difference of opinion as to what was the most important within the different regions. In Kakheti, for example, where drinking water was found to have the highest priority on average across all the communities, only 2.6 percent of respondents prioritized drinking water in Toxliauri and 5.7 percent in Kimiti. By contrast, more than 70 percent prioritized water in Veliscixe. In Mtianeti, where natural gas was given the highest priority, only 3.6 percent prioritized gas in Tskhvarichamia, and virtually no-one voted for gas in Kvesheti. Instead, 62.5 percent voted for electricity in Tskhvarichamia and 81.5 percent voted for telecommunication in Kvesheti. Such a significant intra-regional disparity indicates the danger of generalizing preference at the regional level. The full details of community preferences by community and regions are provided in Annex G. The Drinking Water Sector 99. A significant proportion of the rural population considers that the provision of drinking water is currently inadequate, both in terms of access and quality. Water collection, where necessary, is reported to be both tiring and time consuming. In those locations where centralized supply is available, the quality is perceived to be poor. A number of villagers remarked that they do not drink tap water without treatment because they consider it is unsafe. Local leaders, in all regions except Kakheti, claimed that tap water is not regularly tested. The following text box summarizes some of the more general sentiments: The Problem of Rural Drinking Water As one farmer in Kakheti put it, "one of my family members is always busy with water collection". A local government official in Kvemo Kartli comments; "Water is treated with chloral at the main treatment facility but get polluted because pipeline is damaged. As a result, villagers in this local council claim that if you look into water with glasses, you will think you are looking at herbarium: you can see so many living creatures". 100. Drinking water, however, was given the highest priority only in Kakheti region, and only in Racha region did more than 20 percent of communities vote for drinking water as the priority. This is partly due to the fact that serious water problems can be mitigated by local 42 initiatives: many individual/common wells have been built purely by individual initiatives or under the leadership of local self-governments. Micro-infrastructure projects such as GSIF and Community Mobilization Initiative (CMI) also help communities build wells and water supply system, many of which provide reliable source of water to the great satisfaction of beneficiary communities. 101. Drinking water is given very high priority where, due to unfavorable soil and topographical conditions, digging water is technically difficult. For example, in Ninoshvili, Guria region, where source of water is far distant for half the population and it can take three hours by cart to collect water, as much as 46 percent prioritized drinking water. In almost all regions there exist a number of councils where drinking water is given the highest priority although water is not necessarily a priority in their neighboring councils. 46 percent voted for drinking water in Ninoshvili of Guria region, while the regional average is 12 percent and water is prioritized by less than one percent in five other councils. Also, in Samegrelo, where on average only 4.3 percent voted for drinking water as priority, 38 percent voted for water in Nakifu, where in dry seasons people need to travel one km to fetch water. The Electricity Sector 102. The majority of rural people perceive improvements in the supply of electricity to be the first priority. It is an essential element in running water pumps for drinking water supply, and irrigation systems. The operation of telecommunication systems, mills and other machines necessary in rural life, all require a reliable power supply. The rural population also emphasized that without electricity, they cannot watch television, and so feel cut off from various information sources, and hence opportunities, outside their villages. A problem particularly pronounced for those who live far from the main road networks. Also many pointed out that it is extremely problematic to be living without light during winter. The following text box summarizes some of the hardships: The problems of intermittent and low quality electricity in rural areas One inhabitant in Samegrelo complained that, although power is "supplied" throughout the day, voltage is so low that high voltage items like refrigerators and irons cannot be used until midnight. An inhabitant of Imereti issued a similar sentiment, "...voltage is so low that children have to read under an oil lamp even when power is supplied." According to another inhabitant in Racha, on windy days, the electric wire sways so strongly that it touches structures and short circuits, causing a power surge that damages electric appliances. 103. More than 20 percent of the rural population in the majority of the regions selected electricity as the first priority. The exceptions were Mtianeti and Samtskhe-Javakheti regions. In Guria, where only 72 percent of people are connected with electricity, 42 percent 43 of people prioritized electricity. Intra-regional differences also exist in the demands for the rehabilitation of electricity, though smaller than for other infrastructure. In Mtianeti, no one in Boli and Dzegvi voted for electricity, while 62.5 percent prioritized electricity in Tskhvarichamia. While 23.9 percent voted for electricity in Shida Kartli on average, no one prioritized electricity in Shindisi. 104. Such intra-regional differences reflect, at least partly, groups of villages where small- scale community electrification programs have been implemented, mostly using a hydro power source. For example, in some districts in Kakheti region, local population contributed GEL 30-35 per household to build their own hydro plant which supplies power 24 hours to their satisfaction. Hydro power plants are also constructed under donor-supported community infrastructure projects such as GSIF. Apparently some regions in Georgia are well suited for hydro power generation and hydro plants can be built by local initiative at a low cost. The Irrigation Sector 105. Irrigation was not selected as the highest priority in any of the regions, but about 20 percent voted for irrigation in the regions of Samtskhe-Javakheti and Shida-Kartli. Farmers in Samtskhe-Javakheti region claimed that a lack of irrigation reduced their harvest by as much as 30-40 percent. In some councils of Mtskheta-Tianeti, such as Mukhrani and Nichbisi, the lost output can amount to as much as 70 percent of potential output. Some farmers had stopped producing vegetables and grape. Demand for irrigation is naturally high in such councils, at above 30 percent. In neighboring councils, such as Lisi and Simoniatkhevi, however, no one prioritized irrigation despite the absence of irrigation. As a result, on average only 9.3 percent prioritized irrigation. 106. Less than one percent voted for irrigation in Guria, Racha and Samegrelo regions, where irrigation is unnecessary due to the more favorable climate. Only in two councils, among the 33 surveyed in these regions, irrigation was prioritized by more than two percent, namely, Kvashkhieti and Lia of Samegrelo. Adequate drainage is perceived to be a far more serious problem in these regions. Intra-regional differences also exist for demands for irrigation. People of Kvashkhieti consider that irrigation, currently unnecessary, will be necessary to expand output. In Lia, corn is produced on a relatively large scale (seven ha), and farmers consider that increased production of corn would be possible with irrigation. The Natural Gas Sector 107. Natural gas is given a high priority in many regions, especially Mtianeti, Imereti and Kvemo Kartli, where it is given the highest priority. Many users of natural gas seem to be satisfied with the service, though they consider the tariff levels, and especially the initial cost required to install meters (which can cost as much as GEL 180 ­ 200) too expensive. Those who are not connected wish almost unanimously (85 percent) to be connected. 108. There is a considerable intra-regional disparity in terms of preference for the rehabilitation of the supply of natural gas, which cannot be explained simply by access. For instance, in Shida Kartli, 75.4 percent prioritized gas in Shindisi council, but only 8.1 percent 44 prioritized gas in Teliani councils, although both councils are not currently supplied with gas. State of access and service level, and preference to rehabilitation of gas, does not necessarily coincide in other regions either. It is not clear what combination of factors determines preferences towards the rehabilitation of natural gas, but the availability and price of alternative source of heating and cooking, such as firewood, appear to be important determinants. Rural Georgians prefer natural gas to firewood, reflecting the additional monetary or time costs involved in buying or collecting firewood. However, in the absence of a central heating system, firewood will continue to be used even if gas becomes available, especially during winter to save the overall energy cost, as indicated in the previous section. The Transport Sector 109. The improvement of local roads was ranked as the first priority in three regions, Racha, Samegrelo and Samtskhe-Javakheti, and given priority by more than 15 percent in all regions. Demands for improved roads are especially high in the mountainous Racha region: in Skhvana and Gadishi councils, where more than 70 percent indicated road improvement was the priority. Western and north-western regions tend to have more villages with high preference for road improvement than do eastern regions where less than 20 percent prefer roads in the majority of villages except in a few of Kvemo Kartli region and mountainous councils of Mtianeti region. 110. Farmers in all regions consider that improved roads will help increase income since they are often forced to sell their products to visiting merchants at a very low price, or see their products spoiled, due to a lack of means to transport them to the market. In all regions except for Adjara, more than 40 percent of respondent believe that improved roads will significantly, or slightly, improve their income prospects. In other regions, more than half the surveyed people in Kvemo Kartli and Imereti consider their income will significantly increase. The Telecommunications Sector 111. Improving telecommunications was not given the highest priority in any region, and in no region, except in Mtianeti, Racha and Imereti regions, it was prioritized by more than 10 percent. This is probably due to the availability of mobile phones in many parts of rural Georgia. As mentioned in the previous part, the mobile telephone network covers many parts of rural Georgia, and a lot of the rural population can use the mobile phones of their neighbors or friends at free or low service fees. Since mobile phones are expensive, many in the rural population prefer network telephone and consider its rehabilitation will improve their livelihood by reducing household expenditure, and by increasing the frequency of use of telephone and their exposure to external information and opportunities. 112. Neither mobile nor network telephone is available to very remote areas, and access to telephone service was highly prioritized in such areas. For example, telephone network was prioritized by 81.5 percent in Kvesheti of Mtianeti region, and 68.3 percent in Znakvi of Racha district. These are where communication beyond the close neighborhood becomes almost impossible during bad weather due to very limited access to transport and poor roads, and where it can take as much as three hours to get to where telephone service is accessible. 45 Community preference towards utility ownership 113. Generally speaking, community members prefer state ownership of infrastructure utilities and roads to private ownership. As shown in the Table 12 below, predominant majority of rural population consider state ownership is preferable to other forms of infrastructure ownership.15 There are several reasons for the preference for state ownership. Firstly, rural Georgians tend to consider that government has the responsibility to provide services to citizens, and this is why they pay tax. Secondly they expect government will be more considerate about their welfare than the private sector, and will not just cut service even if they do not pay user fees. Many rural Georgians have trouble paying user fees, and they expect, as some farmers in Racha put it, that while private operators would stop providing services, the public sector would ensure service is provided. By contrast, they do not necessarily consider private ownership to be more efficient or effective than state ownership, since the supply of electricity and gas, both of which are provided by private companies, is no more reliable than supply of water and other services provided by state owned enterprises. Table 12 : Expressed Preferences towards Community Ownership Electricity Gas State owned enterprise 69.6% 80.0% Privately owned company 7.4% 7.9% Local administrative body 4.3% 4.7% Community owned 1.9% 2.3% No preference 16.8% 5.1% 100.0% 100.0% Source: Study data. Potentials of local initiatives and role of local governments 114. The selected priorities do not always coincide with actual preferences, as rehabilitated under local initiatives or donor-financed small-infrastructure projects. This is probably because the infrastructure selected as a priority, by most respondents, requires more than a small-scale intervention, typical of community-based infrastructure projects. Interventions in electricity, roads and natural gas supply require substantial resources which cannot be mobilized locally. Whereas small-scale water supply system or hydro power generation, or roads leading to the cemetery or other communal places, are more suitable to be undertaken locally, frequently under the leadership of the elected mayors. Community members usually contribute most of the cost, but their financial capacity is also very limited. All mayors interviewed mentioned that it is virtually impossible to finance rehabilitation of infrastructure on a local budget. WHAT IS HOUSEHOLD WILLINGNESS TO PAY FOR RURAL INFRASTRUCTURE? 115. The aim of the WTP analysis for households was twofold: 15Preferred ownership was asked only about gas and electricity since other infrastructure, such as water, roads and irrigation, are provided solely by government or owned by community, and respondents will not have solid information to compare state from private ownership. 46 x To estimate the WTP for improved infrastructure services and possibly different levels of the following services: electricity, gas, drinking water, water for irrigation, landline telephone and roads. x To estimate a maximum WTP for infrastructure services. 116. A quantitative approach was undertaken to collect the respondents' explicit choices between defined hypothetical packages of improvements of infrastructure services. The discrete choices were then used to inform an econometric choice model, based on a conditional logit model, to estimate household WTP for various service improvements (Annex J describes the econometric model for estimation of WTP estimates). The usual stages of good design practice were followed: commencing with in-depth interviews, a pilot survey, and then a main survey comprising one thousand (1,000) households (the formal approach employed to select the sample is provided in Annex F). The sample was selected to ensure that the resulting WTP estimates would be representative for the entire population in Georgia. 117. The stated preference questions proposed certain service improvements and commensurate increases in the monthly tariff in the range of GEL 2 to GEL 25 for each of the partial service improvements, and between GEL 6 and GEL 45 for packages of services. The results of this exercise, in absolute terms, are presented in the following table: 47 Table 13 : Willingness to pay estimates for service improvements, by expenditure group, GEL per month Variable Households Households Households with monthly with monthly with monthly expenditures expenditures expenditures between 0-100 between 100- over 200 GEL GEL 200 GEL 8 hours electricity supply, no fluctuations Household with less than 8 h supply 8 8 8 Household with 9-23 hours supply 0 0 0 Household with 24 hours supply, but with voltage fluctuations 0 0 0 WTP for fixed payment, households currently paying by consumption 0 0 0 24 hours electricity supply, no fluctuations Household with less than 8 h supply 12 15 15 Household with 9-23 hours supply 5 5 5 Household with 24 hours supply, but with voltage fluctuations 0 0 0 WTP for fixed payment, households currently paying by consumption 2 2 2 24 h drinking water supply, always safe to drink Households with piped water into dwelling 6 6 6 Households with piped water into yard 11 11 11 Households with own family well 7 7 7 Households using water vendors 25 25 25 Households using public tap/standpipe/well 13 13 13 Households using surface/spring water 14 14 14 Households using water from neighbors 20 20 20 Landline telephone Households with mobile telephone 7 7 7 Households without mobile telephone 3 3 3 Irrigation water during the growing season Households having income from land use, east 7 7 7 Households having income from land use, west 3 3 3 Households not having income from land use, east 4 4 4 Households not having income from land use, west 0 0 0 Improved road All households 4 4 6 Improved roads per minute saved 0,13 0,13 0,13 Source: COWI, (2005). Electricity 118. Households with fewer than eight hours of electricity supply per day have the highest WTP for 24 hours electricity supply (GEL 12-15 per month depending on the level of total household expenditures). These households are also willing to pay GEL 8 per month to get electricity eight hours per day. Whilst the WTP of households who already receive between 9-23 hours of electricity per day is GEL 5 across all income groups. 119. There is also a small difference in the average WTP for 24 hours supply depending on the present method of payment: Households paying per consumed KW/hr currently are 48 willing to pay GEL 2 extra per month in order to pay a fixed fee for electricity 24 hours a day. This can be interpreted either as a preference for paying a fixed monthly amount rather than a varying amount based on actual consumption for budgetary reasons, or it reflects the respondent perception that the fee would be unreflective of consumption. Water supply 120. Households that already have piped water into the dwelling have the lowest WTP to receive 24 hours a day service (GEL 3 per month presumably reflecting current poor quality). Households with water supply in the yard have a slightly higher WTP to receive 24 hour supply into the dwelling (family well: GEL 5 per month; piped water in yard GEL 9 per month), whereas households that need to collect water have a significantly higher WTP: households that buy the water from water vendors have the highest WTP for 24 hours piped water (GEL 31 per month). The WTP does not increase with higher total household expenditures. Landline telephone 121. Households are willing to pay between GEL 5 per month and GEL 8 per month to receive landline telephone connection and unlimited national calls. Households who already have mobile phones are willing to pay more for this service than those without. This reflects that the latter need to communicate and are used to paying a higher rate for their communications. Water for irrigation 122. The WTP for irrigation depends on the geography. Households in the east of Georgia (Kakheti, Shida Kartli, Kvemo Kartli, Samtskhe, Javakheti, Mtskheta Tianeti) have the highest WTP, reflecting the drier climate of those areas. Households in these areas are willing to pay an amount between GEL 6 and GEL 8 per month, and households with income from the use of the land, not surprisingly, have the highest WTP for irrigation water. WTP does not depend on the level of total household expenditures. Households in the west are willing to pay up to GEL 5 per month. Improved roads 123. The WTP for improved roads is estimated as a constant WTP per month and a term that varies with the potential time savings. For instance for households with the lowest household budget, the WTP is calculated: WTP (road improvement, expenditure lower than 200 GEL/month) = 4 GEL + 0.18 GEL * minutes saved per trip 124. Hence, the average WTP for all households varies depending on time saving and the total budget of the household and is presented in Table 14. 49 Table 14 : Willingness to pay estimates for road improvements, by expenditure group, GEL per month (according to the estimated model) Households with monthly Households with monthly expenditures between 0-200 expenditures over 200 GEL, GEL, WTP in GEL per month WTP in GEL per month Expected time saving: 0 4 6 10 5 7 20 6 8 30 8 10 40 9 11 50 10 12 60 12 14 Mean 7 9 Source: COWI (2005). Gas 125. Gas was not included in the stated preference games because the provision of mains gas to households was expected to lead to significant savings in the total household budget, reflecting the lower fuel costs for cooking and heating. The estimated WTP was calculated by assessing the potential savings for households that would follow from the provision of affordable mains gas, based on reported costs of cooking and heating. The expected savings in summer time is GEL 57 per month (sum of expenditures on bottled gas and firewood). In the winter time, the savings are more uncertain as some households appear to use firewood for both cooking and heating, so is estimated at GEL 39 per month. Maximum WTP for infrastructure services 126. The total WTP for all infrastructure services can be estimated in two ways: firstly, all the WTP estimates for the individual infrastructure services can be, and have been, added up. The result is presented in Table 15, and assumes that all households get an improved level of 24 hours electricity supply. The WTP is broadly consistent across all income groups, ranging between GEL 44 and GEL 48 per household per month. Table 15 : Sum of willingness to pay estimates, GEL per month (according to the estimated model) Variable Households with Households with Households with monthly monthly monthly expenditures between expenditures between expenditures over 0-100 GEL 100-200 GEL 200 GEL Sum of WTP 44 46 48 Source: COWI (2005) 127. However, the WTP estimates summarized in this table reflect the ceteris paribus assumption, all other things remaining unchanged. So each WTP amount was obtained from respondents, based on all other infrastructure types remaining the same. Once this principle is sacrificed, i.e. more than one infrastructure type is improved (there is a package of 50 improvements), then the WTP figures `may' exceed actual total WTP for the joint improvement, for the package. A respondent may wish to spend a lower amount on electricity if they know that their water bill is also going to increase due to improved water services. 128. Therefore, the second way of estimating total WTP is by using the information the third stated preference exercise. The results from this exercise can be interpreted as an overall average maximum WTP for infrastructure services. The average total WTP for defined improvements to all services is between GEL 31 and GEL 45 per household per month depending on the expenditure level of the household, cf. Table 16. The WTP for Package 1 is slightly lower than for Package 2 and the WTP for either package depends on the expenditure group. Table 16 : WTP estimates for Package 1 and Package 2, GEL per month (according to the estimated model) Variable Households with Households with Households with monthly monthly monthly expenditures expenditures expenditures over between 0-100 GEL between 100-200 200 GEL GEL Packages Package 1: Electricity and roads 17 22 26 Package 2: Drinking water, irrigation and landline telephone 14 18 19 Total, Package 1 + Package 2 31 39 45 Source: COWI (2005) 129. Comparing the result in Table 16 with the result in Table 13 indicates that the maximum WTP for improvement of all services is lower than the sum of the WTP of the services. For comparison, the two sets of results are reported in the table below along with the ratio between the two estimates of total WTP for all service improvements, called the multiplication factor. The multiplication factor indicates that households in the expenditure group with GEL 0-100 per month in expenditures are willing to pay on average 70 percent of the WTPs given in Table 13 if all services are improved simultaneously. Table 17 : Relation between the sum of WTP and maximum WTP for all services (according to the estimated models) Variable Households with monthly Households with monthly Households with monthly expenditures between 0-100 expenditures between 100- expenditures over 200 GEL GEL 200 GEL Sum of WTP for the services 44 GEL 46 GEL 48 GEL Total maximum WTP 31 GEL 39 GEL 45 GEL Multiplication factor for partial WTPs if all services are 70% 85% 94% improved simultaneously Source: COWI (2005) 51 130. In accordance with the hypothesis, maximum WTP is lower than the sum of the partial WTPs. There is also an income effect such that the sum of partial WTPs is closer to maximum WTP for households with higher total expenditures. The difference between the two estimates is reduced for the two other expenditure groups which accords with the expectation that households in these groups are less budget constrained. Table 17 presents household WTP for the defined service improvements, by expenditure group. WHAT IS ESTIMATED COST TO REHABILITATE THE RURAL INFRASTRUCTURE NETWORKS? The provision of improved roads 131. The required investment necessary to rehabilitate the lifeline or strategic, rural road network, defined as the aggregate of the road sections linking each rural community to the nearest main road, has been calculated in the following way: these main links represent 6,350 km of rural roads, or approximately 40 percent of the total local road network. The road network is in very poor condition overall: out of the roads surveyed, 54 percent are unpaved; and out of the paved roads; and 55 percent are assessed as worn out. The investment for the rehabilitation of the network include: x For unpaved roads: cost of grading and re-gravelling. x For paved roads in very bad condition: pavement rehabilitation works (base recycle together with a double surface treatment DBST). 132. The unit costs of each of the major investment components, based on actual costs, are detailed below. Table 18 : Road interventions and unit costs Road interventions Unit Cost Grading 0.785 US$/m2 Graveling 0.993 US$/m2 Rehabilitation asphalt roads 8.18 US$/ m2 Source: Actual costs from RDMED. 133. Based on this methodology, the total investment for the rehabilitation of the lifeline road network is estimated to be US$ 212 million. Table 19 provides a breakdown of investments by region: 52 Table 19 : Road Investment Cost by region Region Cost (US$) Adjara 13 566 307 Guria 12 772 129 Imereti 35 284 904 Kakheti 26 279 623 Kvemo Kartli 31 061 762 Mtskheta-Mtianeti 24 720 095 Racha-Lechkhumi 5 618 313 Samegrelo 30 907 846 Samtskhe Javakheti 15 707 654 Shida Kartli 16 419 388 TOTAL 212 338 018 The provision of clean drinking water 134. 64 percent of the 1064 communities surveyed in the study are connected to the centralized water supply network. However, the lack of adequate, sometimes any, maintenance has contributed to an alarming deterioration in the distribution network: the pipes are corroded and the pumping stations suffer from an intermittent electric power supply. The consequences are low water pressure, intermittent service, and poor water quality. The residents of affected communities therefore divert to alternative water sources: boreholes, hand dug wells, or untreated open water sources. These are also the sources used for the remaining 36 percent of communities (383), without connection to the centralised distribution network. 135. The cost of rehabilitation of the water supply system, or the extension of such a system to those not currently served, would vary markedly from community to community. It would comprise three elements: a share of the cost of the primary distribution network, which takes the water to the community, the costs of providing/upgrading the secondary distribution network within the community, and finally the costs associated with the actual connections in each dwelling within the community. A robust estimate of the cost would require detailed study at the local level, to capture local variation in the current state of the infrastructure, the density of the settlement, the projected demand, and the topography. 136. However, this information is simply not available currently, so the approach employed here has relied on the use of proxies, to provide an indicative estimate for upgrading/providing connections. The defined proxies are based on regional per capita unit costs, covering the construction/rehabilitation of wells or protected pumps, a water treatment facility, and a household connection where the system is centralized. The resulting estimates of cost for each of the major investment components are detailed below. 53 Table 20 : Water supply investment components and unit costs Investment Component Unit Costs (US$) Protected well/handpump (per capita) 40 Water treatment per community 7,000 Water connection to individual household 50 Source: Study data. 137. The resulting investment cost estimate for rural Georgia is US$47.5 million, based on this approach. However, this estimate does not include the costs associated with the primary or secondary network, which is a major component of total investment cost, but which has to be specifically and carefully estimated for each community. As a comparable estimate, one recent study16, which estimated the cost of upgrading the entire water supply system in Georgia to Millennium Development Goal (MDG) standards, resulted in the figure of US$190 million. Although this figure assumes the replacement of 39 percent of the primary and secondary network. The provision of reliable gas supply 138. The data collected during the course of this study reveals that few rural communities have direct access to the main gas distribution network, with supply limited to a few regional centers. The earlier section identified a number of reasons for the limitations in gas supply, namely failures at a national level (no storage capacity, highly dependant on Russian imports and rapidly increasing costs of energy), and failures at a local level, reflecting the deterioration in the local distribution network, due to a severe lack of maintenance. 139. A review of the survey data allows the categorization of the communities into three broad groups to strengthen the estimated cost of improving provision: A. Communities where gas infrastructure is nonexistent. B. Communities where there is no gas supply but existing gas infrastructure. C. Communities which currently have reliable gas supply. 140. In the case of those communities identified as falling in Group A, the investment cost represents the complete installation of the necessary infrastructure, below the primary network, to ensure provision to the rural community. It includes: (i) the installation of the medium open pressure pipeline, the secondary network, linking the community to the nearest mains gas infrastructure, or primary network; (ii) the installation of the distribution network within the community, the tertiary network; and (iii) the installation of gas meters. Table 21 provides the average unit cost for each component, based on a study in Armenia.17 16COWI (2005) where the costs were calculated using the FEASIBLE toolkit available on the OECD webpage. 17World Bank (2004). 54 Table 21 : Estimated Unit Costs for Installation of New Gas Infrastructure Component Unit cost (US$) Medium pressure pipeline (open) 12,000 -21,000 / km Gas meters 36 / per capita Distribution network 72 / per capita Source: World Bank. 141. In the case of those communities identified as falling in Group B, the rehabilitation costs are estimated as a percentage of the construction costs contained in Table 21, with the proportion reflecting the current condition of the infrastructure. No estimates have been calculated for those communities identified as falling into Group C, reflecting the assumption that only urgent investment costs are included. 142. Table 22 presents estimated total costs for the installation/rehabilitation of the necessary gas infrastructure, broken down by region. The total cost is estimated to amount to US$439 million, although this figure does not consider viability of the investments. Table 22 : Gas distribution rehabilitation costs by region Region Cost (US$) Adjara 33 010 416 Guria 32 974 092 Imereti 57 960 034 Kakheti 29 403 586 Kvemo Kartli 42 382 190 Mtskheta-Mtianeti 18 793 092 Racha-Lechkhumi 57 638 383 Samegrelo 148 444 524 Samtskhe Javakheti 52 529 184 Shida Kartli 17 585 976 TOTAL 490 721 477 Source: Study data. The provision of electricity 143. The electric power sector in Georgia is one of the worst in the countries of the former Soviet Union, with considerable inadequacies in generating capacity, large transmission losses, and an adequate distribution network. In terms of the former, the poorly maintained hydro power plants and thermal power plants function well below their design capacity, at roughly 60 percent, reflecting limitations in both design and maintenance. At the transmission and distribution level, the electric power grid covers a great part of the rural communities of Georgia: on average 96 percent of households are connected to the distribution network. 55 144. However the lack of funds for maintenance and rehabilitation has led to severe degradation of the system and increasing failure rates. The authorities have recently accelerated the reform of the power sector, but the sector is still unable to secure sufficient funds to undertake the necessary rehabilitation. Voltage variations and interruptions remain very frequent, and the number and duration increase during winter. The rural communities surveyed for this study were found to receive electricity for between two and ten hours a day. Households switch to wood and fuel for heating and hot water, increasing indoor pollution. 145. The approach adopted reflects the stylized assumption that the distribution and transmission networks are the primary issues driving costs in the rural areas, rather than the provision of additional generating capacity. This assumption reflects the fact that in nearly all cases, the current distribution network, or grid, is in very poor condition, if not wholly degraded. The provision of additional capacity locally, through mini-hydro stations, is possible in remote areas, but management for local communities remains difficult, and would result in higher unit costs. Generating capacity is seen as a national issue, so the associated costs of adding capacity are excluded for the purposes of this study. 146. The estimated investment costs include the rehabilitation of part of both the transmission (110KV and 35KV) and distribution networks (6KV, 0.4 and 10KV). The cost estimates also include the rehabilitation cost for transmission and substation assets, (breakers, lines), as well as the equipment to protect these assets from further damage. The resulting total cost estimates are presented in Table 23. The costs included in this table were obtained from the United Electric Distribution Company (UED) which covers the whole territory of Georgia except for the Adjara and Kakheti18 regions, and is responsible for 83 percent of total power consumption. 147. The total estimated investment costs amount to US$87 Million. The major cost components are the rehabilitation of the transmission distribution lines and poles and metal towers (78 percent of the costs). The breakdown by region is given below. These estimates were converted in estimates at community level by dividing the regional cost by regional population, and then multiplying the per capita cost by the community population. This is a stylized approach, but in the absence of alternative information was considered to provide a reasonable proxy. Costs for Adjara and Kakheti regions were calculated using UED estimations as proxies. 18Covered respectively by Kakheti Energy Distribution and Adjara distribution. 56 Table 23 : Electric distribution network rehabilitation costs by region Region Investment cost (Million US$) Mtskheta-Mtianeti 6.72 Kvemo Kartli 19.64 Shida Kartli 7.68 Samstkhe Javakheti 5.93 Imereti, Racha Lechkhumi, Kvemo Svaneti 27.49 Samegrelo, Guria, Zemo Svaneti 19.81 Total 87.28 Source: UED The provision of improved irrigation 148. The irrigation system was developed in eastern and central Georgia during former Soviet times. In the last 15-20 years, inadequate maintenance has led to a severe degradation of the infrastructure: out of 370,000 hectares irrigated in 1990, self flowing irrigation systems are functioning only on 150,000 thousand hectares, about 40 percent of the former total. In the case of mechanical elevation systems, the reduction is greater with only 16 percent of the original area still irrigated (12.7 thousand hectares19 out of an original 80 thousand hectares). Although, these latter schemes also suffer from inadequate power supplies. 149. At the level of the rural communities, the most urgent works include the renovation of duckers, aqua duce, tunnels, galleries, pipe side. The rehabilitation also needs to be followed by regular maintenance (cleaning of berm channels by removing soil). Past bank projects have estimated the cost of rehabilitation to be an average of US$800/hectare. The cost of maintenance and exploitation on a self flowing system is estimated by the Department for Amelioration System Management (DASM) at an annual US$22/hectare. 150. The renovation of pumping schemes is more expensive. In addition to the renovation of the cleaning and dissemination channels, there is a need to invest in new pumps stations, the pumps usually being completely depreciated. The maintenance of these pumps, as well as the energy cost, boosts the maintenance and exploitation cost of these mechanical elevation systems. The annual maintenance cost is estimated at US$61/hectare.20 151. The estimated investment costs include only the rehabilitation of formerly properly functioning irrigation schemes. Since irrigation in Georgia has been extensively developed, it can be assumed that irrigation schemes were installed on all land where it is technically suitable, so extension of the irrigated area beyond that originally covered is considered to be unrealistic. Moreover, in evaluating the cost for each surveyed community, the economic justification for rehabilitation was not considered. Indeed, in central Georgia, the renovation of the irrigation system may not be the strategic investments to spur agriculture production. Low quality crops and poor use of fertilizers are also important causes of low yields. A cost 19Department for Amelioration Scheme Management. 20Costs are based on estimated costs provided by DASM. 57 benefit analysis would need to be conducted in each case, before funds for specific irrigation system renovation are provided, particularly for expensive pumping schemes. 152. An average unit cost for irrigation was applied for each community (US$800/hectare). However the cost for each community depends on the topography, the condition, the type of the system in each community. This decentralized data had not been collected by the central department at the time of the study. Local Surveys by NGOs or donors, or past development projects are the only accessible means to have access to specific condition. Based on these assumptions, the total cost for the rehabilitation of irrigation systems is estimated to amount to US$165.8 million. The provision of drainage 153. In western Georgia, inadequate drainage renders fields unusable for agricultural production. As with irrigation systems, drainage systems have suffered from a lack of maintenance. As a result, out of the 115.7 hectares drained in 1987, only 34 hectares subsist today. In some communities, drying networks were completely destroyed on part of draining systems. The investment needed include the rehabilitation trunk channels and main collectors. Maintenance and exploitation cost is evaluated at an annual US$9/hectare. Past bank projects evaluate the average cost of rehabilitation to US$500/hectare. Total cost for rehabilitation in Western Georgia is estimated at 40.85 US$ million. A breakdown by region can be found in Table 24. Again, these estimates were converted in estimates at community level by dividing the regional cost by regional population, and then multiplying the per capita cost by the community population. This is a stylized approach, but in the absence of alternative information was considered to provide a reasonable proxy. Table 24 : Drainage and irrigation investment needs by regions (US$ Millions) Region Cost (US$ Million) Guria (drainage) 10.7 Adjara (drainage) 0.62 Samegrelo (drainage) 29.4 Samtskhe Javakheti 21.2 Shida Kartli 23.8 Kvemo Kartli 42.6 Mtskheta-Mtianeti 11.1 Kakheti 67.5 TOTAL 207 The provision of telecommunications 154. There are two technological options for telecoms access in rural areas; cellular and landlines. The cellular option is a feasible option for the provision of rural telecom services, but are not considered in the study for the following reasons: due to the advanced degree of 58 the liberalization in the sector, the number of private operators and thus investors are high, and regional cellular coverage ranges from 60 percent to 88 percent21 of the country, with 95 percent of the population covered by the end of 2003.22 However the rural population is reluctant to use this technology extensively, because it is perceived to be expensive, compared to landline technology. Most cannot afford to become individual subscribers: they will most likely to continue to rely on shared access schemes, such as those described below. 155. Therefore the estimated investment costs to provide access to the telecommunication network are based on the rehabilitation of landlines, and reflect the following two assumptions: (i) all local infrastructure needs replacement; and (ii) regional centers will be equipped with modern digital systems in the near future. Hence, the investment costs for rural communities have been estimated, with a view to the provision of the telephone connection between rural communities and the regional centers. Table 25 provides the key unit investment costs: Table 25 : Input variables to estimate telecommunications investment costs Description Cost (in US$) Internal transmission system (per subscriber) 50 Digital exchanger (per community): Less than 100 subscribers 4,000 100-250 subscribers 5,000 250-500 subscribers 15,000 500-1,000 subscribers 20,000 Over 1,000 subscribers 25,000 Fiber optic cable (per km) 20,000 MW technology (per community) 70,000 Radio modem technology (per community) 12,000 * Unit costs are based on currently prevailing market prices in Georgia. 156. The necessary investments therefore include: (i) The installation of an external link to regional center [using fiber optic cables linking regional center to community (option 1) or microwave technology (MW) (option 2)]. (ii) The installation of internal network (= internal transmission costs + installation of digital exchanger). 157. In the first case, conventional cables options and radio modem were excluded because it does not allow high volume transmission (necessary for internet access for instance).In addition copper wire requires high maintenance costs. (i) Option 1: fiber optic cable. Rural community exchangers are connected to the main regional exchanger through fiber optic cables installed along power lines for example, or highways (and in the second case they are often already installed 21Geocell 22Georgia Ministry of Foreign Affairs online publication 59 because of the emergency system). The advantages of this technology include low transmission losses, high potential bandwidth, and high security (difficult to tap into). However the drawbacks include high cost when the community is remotely located. (ii) Option 2: microwave technology (MA). Cable link between local and telecom backbone is replaced by point-to-point microwave connections (Wireless fixed access - WiMAX). The absence of cables simplifies the installation and the maintenance of the link, particularly in difficult topographic regions. Moreover, it allows more flexibility in front of growing population areas: customers can be added easily (i.e. without the need to lay new sets of cables). However the signal is very sensible to rain effect and requires a line-of-sight communication. If not, repeaters may be needed, and this increases installation costs. There is also a need to anticipate traffic volume in order to avoid upgrading costs. Hence, a careful thus more expensive design of the installation has to be carried out. The methodology used here only compares, for a given distance, the km-driven unit cost of fiber optic cable to the fixed cost of a microwave link. Thus for a rayon- community inferior to, microwave link will be assumed as more economic than fiber optic cables. 158. In the case of the internal network, the main costs include the internal transmission costs and costs of exchangers. The internal transmission costs are largely fixed for each subscriber (see Table 25) and therefore the total internal transmission costs are higher for densely populated larger communities. Therefore, the investment costs for the telecom infrastructure have been estimated assuming installation of digital exchangers in all of the rural communities. 159. Using the above approach, investment costs for rural Georgia have been estimated at US$43.7 million. The breakdown by region is given in Table 26; Table 26 : Telecommunication investment needs by region Region Cost (US$) Adjara 2 932 600 Guria 2 467 950 Imereti 8 482 800 Kakheti 7 521 550 Kvemo Kartli 5 520 400 Mtskheta-Mtianeti 2 096 650 Racha-Lechkhumi 1 271 500 Samegrelo 6 133 550 Samtskhe Javakheti 3 055 650 Shida Kartli 4 257 450 TOTAL 43 740 100 Source: COWI, (2005) 60 KEY CONCLUSIONS AND NEXT STEPS 160. Inadequate rural infrastructure is a significant contributory factor in the incidence of rural poverty in Georgia. Due to combination of historical factors, opaque institutional arrangements, and lack of donor coordination in the past, the extensive rural infrastructure networks have declined in many areas to the point of complete failure. Addressing the decline of these networks, where economically viable, is seen a key factor in efforts to address rural poverty and improve the lives of rural inhabitants. Despite receiving substantial financing from donors, there has been a limited record of success in earlier interventions. However, in the recent years the coordination of donor-funded activities and the quality of rural infrastructure projects have been improving, and the Government of Georgia should continue this progress. This study investigates the issues of the physical rural infrastructure networks, and access of all rural communities in Georgia to these networks. It provides a complete inventory of transportation, gas, electricity, telecommunications, piped water, and sewerage infrastructure and condition in 1064 rural communities in Georgia. 161. The study has found that the severe degradation of the rural infrastructure networks imposes high economic and social costs on the rural population. For example, 54 percent of rural roads are unpaved and 55 percent of paved roads are in poor or very poor condition, raising costs for users, even when access is possible. Similarly, while almost all rural communities (96 percent) have access to pipe drinking water, only 27 percent of rural households have home taps and water quality is variable and poor. The survey reveals that electricity supply, drinking water and roads/transport are the main priorities selected by majority of communities across the regions. The respondents in the communities expressed significant WTP for improved infrastructure services in their areas of priority. 162. The study presents a GIS database and a `Decision Matrix' methodology, based on a spreadsheet, which together have been entitled the `Rural Infrastructure Assessment Toolkit' (RIAT). This tool has been developed to assist the Government in managing, planning and prioritizing between different infrastructure interventions across sectors, across regions and differing rural communities. Next Steps 163. A broad strategy and detailed action plan are needed for the development of the physical rural infrastructure in Georgia. The intention of this study's findings is to enhance a wider dialogue between the Government, key stakeholders, and the general public on the development of a prospective Rural Infrastructure Strategy. The following are the proposed next steps: 61 Establishment of RIAT in the Road Department. Establish RIAT in the Road Department of the MED (RDMED) of Georgia. This will involve the use of capacity building devices to the Road Department of Georgia and other institutions to operationalize the systems. The following three key elements will be implemented for the sustainable introduction of RIAT: (i) Assessment of the technical aspects of the system: to identify whether it is appropriate for the context. (ii) Technical assistance to facilitate the establishment of the system: to identify required changes for the installment and use of the system. (iii) Training of the operators: to ensure that training is provided to suitable individuals within the stakeholder institutions. Development of the Rural Infrastructure Strategy. Develop Rural Infrastructure Strategy to ensure best use of the limited available resources. The following are the steps in identifying the strategy. (i) Define the investment needs and priorities across sectors, regions, and communities. (ii) Identify critical parts of the network to maintain. (iii) Explore public-private partnership opportunities. (iv) Clarify the issues related to local ownership, maintenance, regulation and standards of infrastructure assets. Discuss the strategy with donors to align government and donor resources within the priorities. Once objectives have been identified the government, particularly the line ministries, and donors, need to coordinate their actions and policies. The government might want to designate an institution to coordinate donor-funded activities. This will in turn lead to better coordination between the government and donors and more consistent and efficient interventions. 164. The draft study has been distributed for comments to key recipient Ministry of Economic Development of Georgia, other governmental and nongovernmental organizations, and donors. Workshop to discuss the submitted comments with participation of stakeholders was held in Tbilisi on September 7, 2006. The final revised version of the study in English and Georgian languages will be distributed soon after the workshop. 62 REFERENCES Binswanger, Hans & Graham Donovan (1987) "Agricultural Modernization - Issues and Options," World Bank Policy Paper Brenneman, A. and M. Kerf, (2002) "Infrastructure and Poverty Linkages: A Literature Review", A Paper Prepared for the Infrastructure Vice Presidency Clarke, George R.G. Scott J. Wallsten, (2002) "Universal (ly Bad) Service: Providing Infrastructure Services to Rural and Poor Urban Consumers", World Bank Policy Research Working Paper 28668, Washington DC. Davey, Kenneth , ed. (2002) "Fiscal Autonomy and Efficiency. Reforms in the Former Soviet Union". OSI, Budapest. DHV. (2004). Study on "Strategy, Regulation and Private Sector Participation in Water Sector of Georgia". Fay, M. & T. Yepes, (2003) "Investing in Infrastructure ­ What is needed from 2000 to 2010?", Policy Research Working Paper 3102, World Bank, Washington DC. GEPLAC ­ Georgian-European Policy and Legal Advice Centre, (2006) Georgian Economic Trends, Quarterly Review, March 2006 Heggie, Ian and Piers Vickers (1998) "Commercial Management and Financing of Roads" Technical Paper 409. Irz, X., Lin, L., Thirtle, C., & S. Wiggins, (2001) "Agricultural Productivity Growth and Poverty Alleviation", Development Policy Review 19(4) pp 449-466. Jalan, J & M Ravallion, (2001) "Does piped water reduce diarrhea for children in rural India?" Policy Reseatch Working Paper 2664 Khandker S.R., Bakht Z., and G.B. Koolwal (2005) "The Poverty Impact of Rural Roads: Evidence from Bangladesh." Mimeo. Lanjouw, Peter & Gershon Feder (2001) "Rural non-fram activities and rural development: from experience towards strategy", Rural Strategy Background Paper Number 4 63 Lebo, J., & D. Schilling, (2001) "Design and Appraisal of Rural Transport Infrastructure ­ Ensuring Basic Access for Rural Communities", World Bank Technical Paper 496. Leipziger, D, Fay, M., Wodon, Q., & Tito Yepes (2003) "Achieving the Millenium Development Goals: The Role of Infrastructure," Policy Research WP 3163 Malmberg-Calvo, C., Ryan, A., & Pouliquen, L., (2001) "Rural Infrastructure Services and Poverty Reduction", A Strategy for Rural Development - Draft, World Bank, Washington D.C. McPherson, K and Bennett, C (2006) "Success Factors for Road Management Systems", World Bank Transport Note 29. Schaeffer, Michael and Makhatadze, Mamuka. (2004) "Revenue and Intergovernmental Assessment and Proposed Action Plan". Unpublished Mimeo. Shergelashvili, Tengiz (2002). "The Financial Aspects of Decentralization", pp.28-43, in Sulukhia, Tamara, Ed. (2002),"Governing City". Tbilisi, Georgia. Songco, Jocelyn A (2002) "Do Rural Infrastructure Investments Benefit the Poor?" Policy Research Working Paper 2796 State Department of Statistics of Georgia/GORBI (2003) "Survey on Rural Settlements' Infrastructure", Tbilisi. Sulukhia, Tamara Ed. (2002),"Governing City". Tbilisi, Georgia. The Government of Georgia, (2003) "Economic Development and Poverty Reduction Strategy Paper", Tbilisi, Georgia. Timmer, C.P., (1997) "How well do the poor connect to the Growth Process", Consulting Assistance on Economic Reform Discussion Paper 178, Institute for International Development, Harvard, Cambridge Massachusetts. Wiggins, Steve and Sharon Procter, (2001) "How Special Are Rural Areas? The Economic Implications of Location for Rural Development", Development Policy Review Vol 19 (4): 427-436 World Bank, (1994) "Infrastructure for Development", Washington DC. World Bank, (1999) "Georgia Poverty and Income Distribution Volume II: Technical Papers", Washington DC. World Bank, (2001) "Utility Pricing and the Poor", Technical Paper no.497, Washington DC. 64 World Bank, (2002) "Georgia Poverty Update", Washington DC. World Bank (2002), PREM, "Moldova Intergovernmental Fiscal Relationships", Draft Paper World Bank, (2003) "Evaluating the Impact of Infrastructure Rehabilitation Projects on Household Welfare in Georgia", Research Working Paper 3155, Washington DC. World Bank (2004a) "Georgia's Secondary and Local Roads Project" Project Appraisal Document. World Bank (2004b) "Georgia Programmatic Poverty Assessment (FY05-07) A Concept Note." Final Draft May 2004 World Bank (2004c) "Measuring Poverty in Georgia ­ Adjusted Methodology" Technical Note ­ Draft World Bank (2006) "Rural infrastructure in Peru: Effectively Underpinning local Development and Fostering Complementarities." January 2006, Report No: 34598-PE 65 APPENDIX 1 ­ THE DECISION MATRIX 165. The Decision Matrix can provide the Government of Georgia with a practical and easy-to-use tool to assist in identifying and prioritizing the most appropriate investments in rural infrastructure. The Decision Matrix identifies a number of criteria, and then employs them in a stylized multi-criteria analysis to ascertain the ranking of the discrete infrastructure interventions. Investments are defined at the community level and by sector (Drinking water, Irrigation and Drainage, Electricity, Roads, Gas, Telecom), and ranked through allocating scores to each intervention in respect of potential impact in terms of a number of defined criteria: poverty alleviation, economic growth, cost effectiveness and community preference. The Decision Matrix is available as an Excel spreadsheet in the attached CD ROM. An example of the Decision Matrix output is shown in Appendix 2 for illustrative purpose. 166. The methodology used was defined to provide the Government of Georgia with a global picture of the rural investments needed, and a Decision Matrix tool to identify how the list of priorities would change, given differing weights on differing criteria, rather than a robust assessment of each potential intervention. A traditional project level assessment for each intervention would be a necessary next step. 167. The criteria included in the Decision Matrix were explicitly selected to reflect both the long and short term strategic objectives of the government, such as poverty reduction, economic growth and effective allocation of scarce public resources. In addition, the Decision Matrix incorporates the preferences of communities thus, to some extent, also addressing affordability and sustainability issues. However, only the first three preferences have been included in the matrix. The number of people registered as unemployed is used as a proxy for poverty due to the paucity of poverty data at the community or rayon level. Finally, the potential for growth was estimated by a number of national experts, using the Delphi method. The intention of the exercise was to use their expertise to make sure that those areas that are considered to have the highest growth potential are represented additionally in the Decision Matrix, to provide the Government with the option to focus interventions in areas of highest economic potential (more information on the approach used is provided in Annex I). The definitions of the criteria used in the Decision Matrix are summarized in the following table: Cost effectiveness Determined using the following formula: [number of people] / [cost of intervention] + [number of registered unemployed] / [cost of intervention]. The number of registered unemployed is used as a proxy to poverty Economic growth Assessed for each rayon by national experts using the Delphi Method (see Annex potential I) Poverty Estimated as the ratio of the population registered unemployed to the total population 66 Cumulative The additive function of community preferences, cost effectiveness and potential ranking for growth Community Determined during the community meetings, represents the percentage of preferences participants who ranked a particular infrastructure first, second and so on. Only the first three preferences are included in the Decision Matrix. 168. Each investment cost was estimated during the study, the methodology used and the detail of these costs is given in chapter on Findings of this study, section on "Estimated Cost to Rehabilitate the Rural Infrastructure". These costs were mostly based on unit costs prevailing in Georgia, collected from state agencies, recent and current World Bank projects, consulting companies. If such data were unavailable, appropriate proxies were identified, based on regional unit costs, and adjusted, as necessary by World Bank or local specialists. Infrastructure condition data was difficult to collect, primarily because of the lack of centralized collection or responsibility for local infrastructure, and the limited capacities to measure at the local level. The investments calculated are aimed to provide MLS, at the lowest cost, for rural communities. Technology alternatives would have to be studied then those mentioned in the study, to take in account the environment and condition of each infrastructure to be rehabilitated, thus ensuring best cost effectiveness. 67 68 MATRIX DECISION THE FROM OUTPUT THE OF EXAMPLE AN ­ 2 APPENDIX