Report No. 40221-CN China Public Services for Building the New Socialist Countryside December 2007 Rural Development, Natural Resources and Environmental Sustainable Development Department East Asia and Pacific Region Document of the World Bank Acknowledgements This report was written by Achim Fock (Task Team Leader; EASRE) and Christine Wong (Intergovernmental Fiscal System; University of Washington), with input from the core team that includes Chaogang Wang (Citizen Participation; EASSD), Chunlin Zhang (Public Services Units; EASPR), Kin Bing Wu (Education; EASHD), Magnus Lindelow (Health; EASHD), Shaojun Li (AAA Coordination; EASRE), and Stuti Khemani (Local Governance; DECRG). Overall guidance was provided by Bert Hofman (Lead Economist, EASPR) and Stephen Mink (Lead Economist; EASRE). The Peer Reviewers were Phil Keefer (DECRG) and Mark George (DFID). The report was edited by Nancy Hearst. This report substantially draws on earlier works by the World Bank and others. These include, in particular, the 2004 World Development Report "Making Services Work for the Poor" and a number of China-specific studies, such as "China Deepening Public Service Unit Reform to Improve Service Delivery" (World Bank 2005a), "China Rural Public Finance for a Harmonious Society" (World Bank 2007a), "Reforming Subnational Finance: Lessons from Northeast China" (World Bank 2007b), and work undertaken to strengthen community participation as well as sectoral studies, for instance that culminating in the report "Health Reform in Rural China: Where Next?" (World Bank 2007c). In addition, the team held meetings with various line agencies at the central level and discussed the topics and issues with national experts. It also undertook two joint field visits with great support from the NDRC; the first visited Chongqing and the second Ganzhou City of Jiangxi. Moreover, a number of background studies were commissioned to fill knowledge gaps. These include two reports on the NDRC's role in public services for rural development, one conducted by Zhang Changchu focusing on social services, and the other by Huang Hanquan focusing on agricultural and infrastructural services. Moreover, Lin Wanlong and Sun Cuiqing made an account of central-level service standards and other policies; of pilots related to public service delivery; and of the information monitored by county-level education and health bureaus. Finally, Liu Mingxing provided an assessment of monitoring and evaluation and local incentives for the provision of education services in rural areas, based on a data set from Gansu Province. We are grateful for the comments by the peer reviewers and others at the conceptual stage of this study and would like to thank Alan Piazza, Christine Qiang, Dan Gibson, Guo Li, Iain Shuker, Shubham Chauduri, and Ulrich Schmitt for providing comments on earlier drafts. Finally, we would like to thank our counterparts at the NDRC for the excellent collaboration and support as well as all other contributors for their participation in this study. Content EXECUTIVE SUMMARY.......................................................................................................... i 1. INTRODUCTION ............................................................................................................ 1 1.1. Background and Objective.................................................................................. 1 1.2. Current Status of Rural Public Services.............................................................. 2 1.3. Framework and Methodology............................................................................. 3 2. CENTRAL (AND PROVINCIAL) GOVERNMENT............................................................... 8 2.1. The Framework and Delegation of Public Service Provision............................. 8 2.2. Financing Public Services................................................................................. 12 2.3. Central-level Monitoring and Evaluation ......................................................... 17 3. LOCAL (COUNTY AND TOWNSHIP) GOVERNMENTS.................................................... 21 3.1. Assignment of Service Delivery....................................................................... 21 3.2. Financing at the Local Government Level........................................................ 23 3.3. Compliance with National Rural Public Service Policies................................. 30 4. SERVICE PROVIDERS.................................................................................................. 34 4.1. PSUs and Delegation by Local Governments................................................... 34 4.2. Financing of PSUs ............................................................................................ 38 4.3. Personnel Management of PSUs....................................................................... 41 4.4. Accountability of Service Providers to Local Governments ............................ 44 5. CLIENTS AND CITIZENS.............................................................................................. 46 5.1. Information for Public Services Provision........................................................ 46 5.2. Consultation and Active Participation in Public Services Provision................ 51 5.3. Financing Public Services................................................................................. 52 5.4. Institutions for Participation ............................................................................. 54 6. SUMMARY AND RECOMMENDATIONS ........................................................................ 58 6.1. Summary........................................................................................................... 58 6.2. Recommendations............................................................................................. 63 7. OUTLINE OF A REFORM PILOT.................................................................................... 69 7.1. Overview........................................................................................................... 69 7.2. Objective and Institutional Set-up .................................................................... 71 7.3. Design Process and Components...................................................................... 71 REFERENCES...................................................................................................................... 76 Annexes Annex 1: Assignment of Responsibility in the Health Sector .......................................... 81 Annex 2: Current Assignments under Local or Joint Financing....................................... 82 Annex 3: Tables on Education Financing......................................................................... 84 Annex 4: Challenge in M&E for performance-based transfers ........................................ 86 Annex 5: Reforming Rural Production and Cultural Services: The Hubei Approach...... 87 Annex 6: Private Management of Public Schools: Examples from Latin America ......... 89 Annex 7: Institutionalizing Participatory Monitoring and Evaluation ............................. 90 Annex 8: CDD Pilots for Rural Poverty Reduction in China........................................... 91 Annex 9: International Experiences of CDD Approaches................................................ 92 Boxes Box 3-1: Perverse incentives of fiscal transfers................................................................ 27 Box 4-1: Delegated and devolved PSUs........................................................................... 35 Box 4-2: Buying public health services in Chongqing and Wuxi .................................... 37 Box 4-3: Price regulation in China's health sector........................................................... 40 Box 5-1: International experience on information campaigns to improve services ......... 47 Box 5-2: Participatory M&E in Ganzhou's New Socialist Countryside campaign.......... 50 Box 5-3: Participatory budgeting in Zeguo Town............................................................ 54 Box 5-4: Pitfalls of creating community organizations.................................................... 56 Box 7-1: Performance-based municipal infrastructure grants in South Africa................. 74 Figures Figure 1-1: Accountability Relationships for Public Service Provision............................. 5 Figure 2-1: Trends in the Breakdown of Central Transfers.............................................. 14 Figure 2-2: Composition of Fiscal Capacity Transfers in China in 1998-2005................ 15 Figure 2-3: Per Capita Transfers by Province in 2005 ..................................................... 16 Figure 2-4: Vertical Control and Accountability in China's Public Health System......... 19 Figure 3-1: Rising Dependency on Transfers in Dayu County......................................... 26 Figure 3-2: Primary School Spending by Province in 2004 ............................................. 30 Figure 4-1: Plot of Average Pay against PSU Own Revenues by Province in 2002........ 40 Figure 5-1: Direct Financing by Citizens.......................................................................... 53 Tables Table 2-1: Key New Programs for Rural Education and Health ...................................... 11 Table 2-2: Trends in Rural (County and Township) Revenues........................................ 13 Table 2-3: The Share of Earmarked Grants to Finance Education ................................... 14 Table 2-4: Expenditure Trends in Rural Compulsory Education ..................................... 16 Table 2-5: Differing Reports of Budgetary Expenditure on Education............................ 18 Table 3-1: Trends in Shares of Rural Public Finance....................................................... 23 Table 3-2: Central Government Ceilings for TEOS Norms ............................................. 29 Table 3-3: Trends in Education Spending in Dayu County.............................................. 32 Table 4-1: Overview of Health Care Providers ................................................................ 35 Table 4-2: Shares of PSU Financing in Budgets at Different Levels in 2002 .................. 38 Table 4-3: The Revenue Status of PSUs in 2002.............................................................. 39 Table 5-1: Comparison between Conventional and Participatory M&E.......................... 49 Table 5-2: Community Participation at Different Stages in Different Sectors................. 55 EXECUTIVE SUMMARY Improving the delivery of public services to the rural populace is a core component of China's new development strategy of building a Harmonious Society and a "New Socialist Countryside." Public services such as education, health care, and social protection are essential to efforts to reduce the gap between urban and rural residents by enabling rural citizens to acquire the human capital needed for participation in China's economic success. The central government's support to rural services has increased substantially in recent years, bringing significant improvements. In addition to increasing transfers, the central government has broadened the scope and raised the minimum standards for services that local governments should deliver. Increasingly, this is done through rolling out new programs ranging from investments in rural infrastructure to rural education, the New Cooperative Medical Scheme, and rural dibao. An increased central government role is essential as part of the Harmonious Society program, given the large regional income differences in China. At present, regional disparities are very large in public services under decentralized finance, reflecting the large income and fiscal disparities across localities. By imposing a national framework that sets minimum standards for many public services and providing funding support, the central government is helping to promote greater uniformity in the delivery of services across the nation. To further improve rural public services, the government faces formidable challenges. First, the financing needs for improving rural public services are huge, given the enormous size of the rural population and the current large "deficits" in services and service quality in rural areas. More importantly, increasing funding for rural services is not straightforward. Under China's highly decentralized systems of fiscal management and administration, the central government employs extensive delegation of authorities and responsibilities in a nested, hierarchical setting. The central government delegates authorities to the provinces and depends on the provinces to carry out their responsibilities. The provinces in turn delegate to the municipalities and depend on them to "deliver" on their assigned responsibilities, and so on downward through the hierarchy. This decentralized administration can be an asset for cost-effective service delivery if local governments can be held accountable for performance. At present, though, the accountability relationships are weak at multiple levels, with the result that compliance with central policies is not always assured at the local levels and central transfers are not always used as intended. Also contributing to the often low effectiveness of public expenditures is the weak accountability relationship between service providers ­ which are mostly public institutions ­ and local governments; and the limited involvement of citizen in public service planning, provision, and monitoring. Just as the central government has few levers to enforce compliance by local governments, local governments often lack effective levers over public service units (PSUs). Finally, the "downward" accountability of these service providers and local governments to citizens is also weak and, despite i recent improvements, most services are provided without significant participation by citizens or communities. This report aims to assist the central government to design a strategy for improving public service delivery in rural areas, with a pilot to fast-track implementation of specific aspects of this strategy. In assessing current public service delivery in rural China, the report adopts a methodological framework that focuses on the accountability relationships among the four key stakeholders: the central government, local governments, service providers, and rural citizens.1 Looking forward, to strengthen accountability in these relationships, reforms must address the following four closely interrelated areas: 1. Appropriate and clearly specified responsibilities The current assignment of responsibilities across subnational governments is murky ­ the Budget Law (1994) delineates only between central and "local" governments, and leaves it to subnational governments to work out how to divide the "local" responsibilities. This clouds the accountability relationship because, without a formal system of responsibilities for each level of the subnational government, local governments do not have clearly articulated roles and functions against which they can be held accountable. Making explicit and formal the responsibilities of each level of government is thus a precondition for strengthening accountability between the various levels of government. Similarly, PSUs should be given clear assignments of responsibilities, through the further strengthening of the use of "letters of responsibilities," "education objective- accountability agreements," and other performance compacts that specify tasks to be performed. The boundaries between government and PSUs should be more clearly set and the current practice of PSUs performing core government or administrative functions should be phased out. Finally, the role of citizens should be further enhanced, and rights and practices to participation by citizens or communities in the design, delivery, and monitoring and evaluation of public service delivery should be strengthened. 2. Financing Inadequate finance remains a fundamental obstacle to the intergovernmental accountability relationship today. Despite significantly increased transfers, the intergovernmental fiscal system still does not ensure sufficient funding to counties and townships in the western and central provinces. This is true for overall amounts as well as for specific programs. Furthermore, the intergovernmental fiscal system still lacks a mechanism for ensuring that mandated services can be financed in poor counties. Under-resourcing must be eliminated since it has a significant detrimental effect on monitoring and evaluation ­ if tasks exceed resources, it is hard to judge performance failures or to hold local governments accountable for them. Likewise, PSUs must be financed adequately for their public service tasks, whether through budgetary appropriations, (regulated) user charges, or other subventions. 1This framework is drawn from World Bank (2004a). ii 3. Information The information base for policy analysis in China is weak, especially for rural public services, and this undermines efforts to judge the performance of local governments and service providers. Even though a huge amount of information is routinely collected, China lacks a system for vetting and reconciling the data reported by the different ministries and agencies, and their figures can vary widely. Discrepancies sometimes exist even for financial data reported by different departments of the same ministry. Data on social indicators and service outputs are weaker still ­ figures for school enrollments, hospital bed usage, etc. are widely considered unreliable. Importantly for rural services, because "rural" is not a statistical reporting category, estimates of the "rural" share are typically based on administrative divisions and are necessarily coarse.2 Building a system of reporting on service delivery - both upward and downward - will be vital to monitoring and evaluation. Accurate and timely information is essential for better policy-making as well as for citizen and community participation toward improved services. Continued efforts should be made towards more transparency and the direct involvement of citizens and their user groups and civil society organizations in the monitoring and evaluation of public services. The central government should take the lead in improving data harmonization, sharing, and transparent commissioning of evaluations of its policies and programs. 4. Enforcement of responsibilities Enforcement of public service delivery often is not clearly assigned to any level of government3 and relies heavily on the personal responsibility system using the performance reviews of government officials. The system of personal responsibility has been effective in the past to enforce selected objectives such as economic development and family planning, and service-orientation is being included in the performance evaluations of local officials in some pilot reforms. However, sole reliance on the personal responsibility system has its limitations and the central government should make greater use of transfers as financial incentives to induce the compliance of local governments in delivering rural services, e.g., in the form of more transfers or more flexibility in the use of funding resources. Similarly, local governments could use financial incentives for service providers to stimulate competition. However, no enforcement mechanism can overcome systemic constraints, and improving enforcement will require fixing the other shortcomings discussed above. In sum, strengthening the accountability relationships among the key stakeholders will require: clarifying and specifying responsibilities for each level of government and for service providers, 2 In education, for example, "rural" includes all county- and township-level institutions. As urbanization claims an increasing proportion of county-level units, and county towns take on increasingly "urban" characteristics, these estimates will tend to inflate the "rural" share of expenditures and underestimate the urban-rural disparities. 3Exceptions are made during national campaigns, such as the campaign to universalize nine years of rural compulsory education. Under this campaign, responsibility is assigned to the provinces and counties. iii ensuring adequate financing, building a robust information reporting and evaluation system, and creating mechanisms for effectively enforcing responsibilities. These reforms will be the essential building blocks for the effective management of a vast, decentralized economy, where the central government can delegate responsibilities to lower levels and have the wherewithal to ensure that policies are implemented. Until these building blocks are in place, central government efforts to improve rural public services will continue to be stymied by the unreliable transmission of policies and resources downward through the multiple levels. Outline of a Pilot Reform Based on the above recommendations, this report sketches out a pilot reform that (i) tackles several of the shortcomings identified; (ii) is sufficiently comprehensive in its design to effectively build on and be consistent with other reform programs; and (iii) is sufficiently ambitious to have a substantial impact on service delivery, but can be tested in a limited area. At the core of the proposed reform pilot is the strengthening of two interlinked factors that provide incentives to county governments: fiscal transfers for improved service delivery and the monitoring and evaluation system. The objective of the reform pilot is to contribute to improving the delivery of public goods and services in rural areas, and, more specifically, to achieve minimum standards of service delivery in poor rural areas. Its success would be measured in terms of results or impacts (e.g., school enrollments, drop-out rates, quality of learning, etc.) and cost-effectiveness. In line with China's positive experiences with pilots, it would be tested on a limited scale, in a limited number of counties within only one or very few provinces (in central or western China). A political commitment to such reform will be the key selection criteria for participating localities. The reform pilot is a process that needs to be well planned and implemented in steps over a number of years. It would include three components. The existing (i) monitoring and evaluation system for the provision of public goods would be strengthened; the pilot counties would then be given (ii) financial incentives, either in the form of sectoral block grants to replace existing earmarked transfers or through additional funding, and (iii) training and technical assistance in a broad range of areas of public service delivery. The counties' performance in rural public service provision would determine the financial incentives and other forms of recognition they receive. Although the scope of the pilot should be as comprehensive as possible, a sectoral focus is possible. iv 1. INTRODUCTION 1.1. Background and Objective Entering the 21st century, China has adopted a new development paradigm that emphasizes the building of a Harmonious Society with more balanced development across regions and across sectors. 4 Under this new development paradigm, the government has substantially increased its commitment to pro-poor, pro-rural programs. This is reflected in numerous official statements,5 and the many new programs introduced over the past few years. Building a "New Socialist Countryside" is a key part of the promise to "extract less, put more back in, and enliven" the rural sector. This is a timely shift in policy to redress the large disparities that have emerged across sectors and regions in the course of China's remarkable economic growth, especially over the past decade.6 The greatest disparities are those between urban and rural residents. By virtually all estimates, the average urban income per capita is now more than three times that of the average rural income per capita, a gap that is among the largest in the world. In addition, rural citizens enjoy public services that are far inferior to those provided to their urban counterparts. Reducing these differences is a critical step toward building a Harmonious Society that includes all citizens. In implementing this new development paradigm, the government faces formidable challenges. First, the financing needs for improving rural public services are huge, given the enormous size of the rural population and the existing large "deficits" in services and service quality in rural areas. Bringing rural services up to the level of urban services would require an enormous commitment of resources.7 Second, in China's highly decentralized fiscal and administrative systems (explained later in the report), the vast majority of public services are provided and financed by local governments. Since the central government accounts for less than 10 percent of budgetary expenditures on social services, such as education, health, and social relief (Table 1-1), its control over social outcomes in the rural sector ­ e.g., to improve rural public services ­ is at best attenuated, and compliance with central policies is not always assured at the local levels. Even when the central government injects resources to support local services, these resources pass through the provinces and municipalities before reaching the counties. Leakages can occur at each level, including the county, and at present the central government does not have sufficient few levers for holding local governments accountable. 4The paradigm adopts a "scientific view of the development process" (kexue fazhan guan) that emphasizes sustainable growth and "putting people first" (yiren weiben). This is laid out in some detail in the Eleventh Five-Year Plan and also explained in Wen Jiabao, "Firmly Establish and Resolutely Implement a Scientific Development View," his closing speech at the specialized research course for provincial-level cadres on establishing and implementing a scientific development view, February 21, 2004. 5Since 2003, the first document issued by the State Council each year, popularly known as the "Number One Document" where the top emphasis of the government is laid out, has been devoted to rural issues. 6For recent studies of these disparities, see Ravallion and Chen (2004), and World Bank (2007a). 7For example, at the National Conference on Health in January 2007, both Hu Jintao and Wen Jiabao called for building a health care system that "covers both the urban and rural populace" (Ministry of Health Web site January 30, 2007). 1 Table 1-1: The Division of Budgetary Expenditures on Major Public Services (2005) Share of Proportion of total budget component Central Sub- Nation. Central Sub- national Budget national Expenditures for Capital Construction 39.1 60.9 10.1 15.3 8.3 Operating Expenses for Education 6.2 93.8 11.6 2.8 14.7 Operating Expenses for Health 2.1 97.9 3.0 0.2 4.0 Social Welfare and Relief 0.7 99.3 2.1 0.1 2.8 Total expenditures* 25.8 74.2 *Total expenditures are adjusted to account for enterprise loss subsidies. Source: CSY (2006); CFY (2005). This report is designed to assist the central government, and especially the NDRC, in defining an overall strategy for improving public service delivery in rural areas. In addition, it aims to assist the NDRC and related agencies in defining a specific reform for fast-track implementation of some specific aspects of this strategy. To this end, the study assesses current public service delivery in rural China and provides policy recommendations to strengthen not only policies and programs, but also reforms of the allocation mechanisms and management systems in support of building the New Socialist Countryside. 1.2. Current Status of Rural Public Services The urban-rural gap in public services is large, and in poor regions the gap often falls short of national policy stipulations. The estimate of China's human development index for 2003 was 0.81 for urban areas and only 0.67 for rural areas (UNDP 2005). Apart from the differences in income, this indicator reflects the lower life expectancy in rural areas (which, at 69.6, was 5.6 years less than that in urban areas), as well as differences in level of education. The share of the population between 15 to 64 years of age without any formal education was 8.7 percent in rural areas ­ more than three times the urban rate. Rural-urban differences in service delivery are also reflected in funding, access, and quality. For example, the per capita recurrent expenditure on education was three times higher in city districts than in rural communities; and health expenditures were almost three times as high in urban areas compared to those in rural areas (2003 figures; UNDP 2005). In terms of access to services, just one example might suffice: The number of doctors per 1000 persons was 5.2 in urban areas, but only 2.7 in rural areas (2002 figures; Ministry of Health 2004). Some of the differences in outcomes are stark: infant mortality (2001) was nearly three times as high at 33.8 (per 1000 live births), compared to the urban rate of 13.6. Similar urban-rural differences exist with respect to education and infrastructural services such as water, sanitation, roads, and information and communication technology (ICT).8 Substantial efforts are being made to address the challenges of improving rural public services. Major programs to improve public service delivery have been rolled out over the last few years, including programs for the rural areas. Public spending for rural 8A substantial share of the increasing inequality, including within the rural areas, can be traced to differences in human capital (World Bank 2007d). 2 public services is increasing; major improvements are being made to ensure salary payments at the local government level; administrative reforms are taking place at the county, township, and village levels; and reforms of public services units (PSUs) are being piloted in many parts of the country. New sectoral initiatives are being introduced, such as the free Rural Compulsory Education program, the New Rural Cooperative Medical Scheme (NCMS), the Rural Medical Assistance Scheme, and the rural minimum living stipend (dibao). Improving the effectiveness of the new programs will require better coordination and strengthening the accountability of local governments and service providers. To date, implementation of many programs has been uneven, and there are signs that the benefits of reform are sometimes less than expected. One key difficulty is that central and local interests do not always coincide, so that local governments may assign lower weight to some program objectives that are central government priorities. For example, reductions in out-of-pocket costs for medical care have been small under the NCMS because county governments concerned about the financial costs to the county have tended to produce insurance schemes with conservative designs. With some reforms working at cross-purposes, there have been some negative effects on rural public service delivery. Most noticeably, the Rural-Fee-Reform and the subsequent elimination of agricultural taxes ­ while overall positive due to the benefits they bring to farmers ­ have reduced local governments' own revenues for service delivery. 1.3. Framework and Methodology This report adopts as its framework for assessing public service delivery the accountability triangle involving policy-makers, service providers, and citizens. This framework was presented in the World Development Report 2004, and centers on the accountability of these stakeholders to one another. For the purpose of this study, the framework is expanded by separating "policy-makers" into central (higher-level) government and local (lower-level) government. For almost all rural public services, "local government" is defined as the county and township levels. The report uses this framework to look at the role of the four key stakeholders and their relationships to one another: central (and provincial) government ­ county and township governments ­ PSUs and other service providers ­ villages and rural citizens.9 Taking this approach is not simple for the case of China, a country in transition featuring rapid economic and political development, a changing role of the state, and large differences across regions. In fact, one of the difficulties of public service provision in China ­ as well as in conducting this analysis ­ lies in the fuzzy delineation of responsibilities between various levels of government and between local governments and service providers, which vary considerably over space and time. Moreover, villages assume the role of self-governing bodies representing the clients of rural public services, but are also the extended arms of the government in implementing policies, including service provision. Nevertheless, using the basic framework as a point of reference allows 9Obviously, this is a crude way of portraying the Chinese "system." Central and provincial governments have different interests; the relationship between counties and townships is crucial for the delivery of rural public services; prefectures/cities are not even mentioned in this framework; etc. Although this report will point to these issues, it often abstracts from these "details" for the sake of simplicity. 3 identification of those issues facing each major stakeholder and discussion of potential associated problems for service delivery. The relationship between the individual stakeholders is driven by interests or incentives, which are substantially determined by the other stakeholders, while also shaped by institutions (laws, regulations, traditions, etc.). The strength of the accountability relationships between these different stakeholders determines the service delivery system and its outcomes. In the terminology of the 2004 World Development Report, accountability is the relationship among actors that has five features. The "principal" (i) delegates and (ii) provides the necessary financing to the "agent," i.e., accountable actor, who then (iii) performs the service. The "principal" also (iv) collects and analyzes information, and (v) enforces the accountability relationship by affecting the interest of the "agent."10 The report looks at the role of the various stakeholders from the perspective of the central government, the target audience for this report. The central government sets the national policy objectives for public service delivery and determines ­ through direct or indirect measures ­ the roles of the other key stakeholders in public service delivery and their relationship to one another. As illustrated in Figure 1-1, the central government has a direct accountability relationship with the local governments. In addition, it sets the regulatory framework for the roles that service providers, in particular PSUs, and citizens assume in the provision of rural public services. Although all three "local stakeholders" have their own interests, these interests or incentives are to a considerable extent set by the central government's rules and regulations, financing, monitoring and evaluation, and rewards and punishments. We first examine the "upward" accountability relationship between the central and local governments. In Chapter 2 we assess the role of the central government and the instruments it uses to implement national policies. We put particular emphasis on how the central government delegates the provision of rural public services to the different levels of local governments, relying on provincial governments as a key intermediary (or in loco parentis). For rural public services, the central government retains control by setting standards as well as using regulations to define the scope and flexibility of local governments. We further look at the intergovernmental fiscal system, which determines the amount and type of financing local governments have available for the provision of services; this requires a look at overall revenue assignments as well as general or earmarked transfers from the central and provincial to the county and township levels. Finally, we look at the monitoring and evaluation system that the central government has put in place to gather and analyze information on service delivery by local governments and how it relates to the various programs and policies set by the central government. 10To take this accountability relationship to an "institutional" level, i.e., between central and local governments, service providers, and citizens, might seem like a "foreign concept" in the Chinese context, and we believe it is. The Chinese system has been largely relying ­ very successfully ­ on the strong personal responsibilities of public employees through the government system and service provider, and under the leadership of the Party and the central government. Given our comparative advantage, we intend to complement this system rather than to question it as a successful model for China's current stage of development. 4 Figure 1-1: Accountability Relationships for Public Service Provision Higher-level Framework / Regulations Policy-makers for Service Providers Framework / Regulations Upward accountability for (objectives/targets; monitoring; fiscal transfers; appointments) Participation Local Policy-makers Upward accountability Downward accountability (targets; financing; monitoring; appointments) ("state power"; consultation) Service Providers Rural Citizens / Direct downward accountability Public Service Users (participation) Source: Adopted from World Bank (2004a). Chapter 3 completes the analysis of the central-local accountability relationship by analyzing the incentive structure of local (county and township) governments. This chapter takes the perspective of local governments, which are responsible for "performance," i.e., for implementing the service standards set by the central government given available financial resources and for reporting on performance to the higher levels. As we will show, county and township governments have substantial freedom to decide on how to implement national or provincial standards, complementing them with additional regulations, topping up expenditures with own fiscal resources, etc. Given this context, the framework set by the higher levels influences the local governments' capacity and incentives to perform, especially in the dynamic environment of the past few years when both policies and fiscal resource flows were changing rapidly. In particular, we assess how the central and provincial governments' enforcement mechanisms are impacting the local levels.11 Chapter 4 analyzes the role of the service providers and their upward accountability relationship with local governments. This includes local governments' delegation of responsibilities; the finances made available to service providers; the personnel system, another very important part of incentives and enforcement at this level; the monitoring and evaluation; and enforcement. Importantly, we also focus on assessing the impact that the central (and provincial) government has ­ mainly through regulations ­ on the role of service providers and their relationship with local governments through central regulations for these units (and the enforcement of these regulations). Chapter 5 looks at the role of rural citizens as a fourth stakeholder in the provision of local services, and at their role and relationships with the local government and 11These various elements of, on the one hand, delegation, financing, and monitoring and, on the other, enforcement are not independent from one another. 5 service providers. This analysis focuses on the process through which the demand for public services is voiced as well as on the degree and level of participation by rural citizens in service provision, through information, consultation, or direct involvement in the provision of public services. Citizens also play a role in the financing of public goods, through influencing the allocative decisions of local governments and schools and through own funding of services. Finally, they have an important role in the monitoring of services. The special role of villages and other community organizations representing citizens is also discussed in this chapter. Again, particular emphasis is put on the central (and provincial) governments' role in setting the framework for citizens and their communities (and in enforcing these rules). Chapter 6 summarizes the analysis and draw conclusions for the strengthening of public service provision in rural areas. It tries to focus on the key areas of central government action that would lead to a more effective institutional set-up. Finally, Chapter 7 outlines a pilot reform proposal. The reform would address some of the key issues identified as impeding more efficient public service delivery. Scope Rural public services. The study takes a "spatial" definition of rural public services, i.e., rather than focusing on public services to all citizens with a rural hukou, it focuses on public services for citizens living in rural areas.12 The focus is on public services provided at the county, township, and village levels, which can be defined as rural areas, and where the citizens living in these locations also access the services. The analysis of the county will be central in this study because it has the greatest responsibilities for most rural public services13 and a number of services for rural citizens are also provided at that level, for example through the Center for Disease Control (CDC). Public Services. In this report we adopt a definition for public services as "any service provided for large numbers of citizens, in which there is potential significant market failure (broadly interpreted to include equity as well as efficiency) justifying government involvement ­ whether in production, finance, or regulation" (Grout and Stevens 2003). Sectoral scope. The study, by relying on expertise from a range of sectors, aims at assessing factors and providing recommendations on issues of service delivery that are cross-cutting to most, if not all, public services, including intergovernmental fiscal relations, performance management and accountability across levels of government, the role of government in financing, regulating, and providing public services, and the 12There is a substantial discussion among policy-makers and academicians as to the impact of migration on public service delivery. Half or more of the working age population are often not living and working in the villages where they have hukou, so this clearly is a topic of tremendous importance. However, this study will not be able to provide answers as to where the government should best provide public services ­ in rural areas where people are registered and where services are generally cheaper (also because they are of a lower standard than those in urban areas) or in urban areas so as to better integrate migrant workers. What seems clear is that improvements should take place at both ends (see World Bank Policy Note on Employment of Migrants) and improvements in education and other public services are also key to a successful migration policy. 13In fact, the role of the county has become more central as responsibilities for rural public service provision, including in key areas such as health care, education, or social safety, have increasingly moved up from the villages and townships to this level. 6 organization and financing of service providers. Where appropriate, the report will draw on information on key social services, especially education and health, to provide illustrative examples.14 Regional scope. The level and quality of service delivery vary tremendously across the country ­ a reflection of the vast differences in geographical, socio-economic, and political factors across the country. To some extent, the study has to account for these differences as they require a differentiated approach by the central (and provincial) governments to work with and impact the local (grassroots) governments and other stakeholders in improving service delivery. However, given the limited time and resources for undertaking the study, a generalization of issues and recommendations is unavoidable; a focus on the poorer inland areas that are more dependent on higher-level support is thus a rational choice for this study. 14According to the 2006 Document Number 1, this includes rural nine-year compulsory education, skills training for rural labor, rural health, rural "cultural development," and the establishment of a rural social security system (including social relief [wubao, dibao], rural social pensions, family planning, and poverty reduction). Given the time and resources available for the study, in-depth investigations of any particular sector cannot be achieved. The study can, however, rely on recent (or ongoing) sectoral studies, such as the study on rural health (World Bank 2006b). 7 2. CENTRAL (AND PROVINCIAL) GOVERNMENT This chapter and the next (Chapter 3) will together demonstrate that even though they have increased, transfers by the central government to poor counties have not kept pace with the increased responsibilities assigned to counties. At the same time, the incentive framework for disbursing these funds to meet local service delivery needs is weak. The weakness corresponds to three shortcomings: central and provincial government officials have the authority to hold local government officials accountable for poor performance, but lack performance data essential for accountability; the severe mismatch of responsibilities and resources makes the assessment of performance very difficult (see below for more on this); and service delivery beneficiaries, with first-hand information on service quality, lack the means to hold local government officials accountable, the ability to transmit their information on quality up to higher levels of government, and any knowledge of performance generally that would allow them to benchmark performance by their own county officials. 2.1. The Framework and Delegation of Public Service Provision To implement national policies, China employs a system of extensive delegation of authorities and responsibilities, whereby the central government delegates authorities to the provinces and depends on the provinces to carry out their responsibilities. The provinces in turn delegate to the municipalities and depend on them to "deliver" on their assigned responsibilities, and so on downward through the hierarchy. This decentralized administration can be an asset for cost-effective service delivery, if lower-level agents can be held accountable for performance. At present, these accountability relationships are weak at all four key levels: that between the central and provincial governments, provincial and local governments, local governments and service providers, and service providers and citizens. The central government in China plays a key role in setting the national development agenda. The central government's vision is developed in the Five-Year Plans, and specific programs and investments are announced in the reports on the work of the government presented to the National People's Congress by the Premier each year in March. To achieve the national development agenda and to maintain some uniformity of services provided, the central government sets the policy framework for the provision of public services and stipulates minimum standards for services that are considered to be national priorities. At present, this list includes access to compulsory education, preventing the spread of infectious diseases, and the provision of assistance in health care and income support for the poorest citizens. For these key services, central line ministries develop guidelines for required inputs, such as student-teacher ratios, a township health center in every township, etc.15 15The central government also sets the framework for the structure and organization of service providers as well as for clients and citizens in the provision of public services. These aspects are discussed in Chapters 4 and 5. 8 In many countries, the assignment of responsibilities to local governments is formalized in law (such as the Local Government Code in the Philippines), or is supervised by a formal body, such as the Ministry of Internal Affairs and Communications in Japan,16 or the Council of Australian Governments in Australia. In Germany, under the system of "cooperative federalism," decisions are coordinated through an extensive net of multilevel committees.17 In China, which is a unitary country, there are no formal institutions that govern decentralization. The Constitution formally assigns all powers to the central government, including the authority to assign and adjust responsibilities at lower levels.18 The administrative structure of government in China is highly decentralized, with extensive delegation of authorities and responsibilities downward in a nested hierarchy. This system of extensive delegation is managed by a central bureaucracy that is exceptionally small ­ in core departments of only some 50,000 civil servants, who sit atop a bureaucratic pyramid that has more than 32 million public employees. The central government assigns authorities and responsibilities to the provinces, and relies on the provinces to "deliver" the national targets, e.g., to universalize basic education, to promote the spread of the New Rural Cooperative Medical Scheme, etc. In turn, the provinces make assignments to the prefectures and depend on them to implement policies within their territories, and the prefectures turn to the counties, and so on. This hierarchical management of decentralization is built, albeit implicitly, into the Budget Law (1994), which spells out the division of responsibilities between the central and subnational governments, but leaves it to the provinces to work out the details of division with lower levels of government. Higher-level authorities formulate broad policies and guidelines for the work of the lower levels, but finance institutions only at their own level, while providing limited funding for the next lower level. This system of hierarchical management is illustrated in Annex 1, which presents the assignment of responsibilities in the health sector. It is clear from the table in the annex that the governments at the county and township levels are responsible for financing, delivering, and managing both health care and public health services for the rural populace. The assignments are broadly similar for education and social welfare services, whereby rural governments at the grassroots levels bear heavy financial and administrative responsibilities. As will be shown below, this "traditional" administrative system is undergoing reform. In recent years, to support the many new programs and initiatives designed to advance the Harmonious Society agenda, the central government has introduced a number of changes aimed at streamlining the administrative structure ­ by reducing the number of tiers involved in rural policy implementation, reducing the number of administrative units at each level, and concentrating management responsibilities at three levels: the central government, provinces, and counties. 16Until 2001 it was called the Ministry of Home Affairs. 17For references on how decentralization is managed in other countries, see World Bank (2005), Muramatsu et al. (2001), and Martinez-Vazquez and Searle (2006). 18The Constitution is available at http://english.peopledaily.com.cn/constitution/constitution.html 9 In this system, provinces play a critical role in both transmitting and adapting central government policies to fit conditions for local implementation ­ to some extent filling the distance between central and local governments. The provinces formulate expenditure standards, oversee implementation, and collect statistics for upward reporting. The provincial role will be further strengthened under the current reforms to streamline the administrative structure. However, at present, the role of provinces is not specified by law or decree and provincial officials have significant latitude in deciding how to pass on central policies and resources (Li and Xu 2006). As some provinces play a far more active role than others in ensuring the implementation of national policies ­ including in the redistribution of fiscal resources - the outcome is that significant differences exist both in the assignment of responsibilities and in fiscal disparities among local governments. Empirical studies have found clear "provincial effects" due to the differences in provincial policy (World Bank 2007b). Municipalities and prefectures have similar functions and add another layer. They also interpret higher-level mandates for local-level implementation, channel and partially add to financial resources from above, and play a central role in information flow and ensuring compliance. Reforms piloted in a number of provinces put counties directly under the fiscal management of provinces. For example, in Jiangxi Province piloting for this reform began in 2005 and will be progressively expanded. By removing municipalities from the hierarchical management of rural governments, the reform aims to improve efficiency and the flow of resources for rural development. In the process, it is likely to increase the powers and authorities vested in the county. However, the current reform affects only the downward flow of fiscal resources and not the other responsibilities of municipalities, including for rural development and the delivery of public services. It also does not eliminate the responsibility of municipalities to contribute to the sharing of the "local" portion of funding for new programs, such as the TEOS and free rural basic education. Consequently, the role of the municipality/prefecture, though varying significantly across the country, will continue to be very important. To support achievement of national goals and ensure that services across the nation meet some minimum quality standards, the central government (often through its line ministries) issues policies and regulations that establish requirements for key sectors and services. In education, for example, qualification standards are established for teachers at each level, along with recommended student-teacher ratios, floor space and equipment specifications for schools, etc. In the health sector, central policies require that, inter alia: - Key public health programs and policies should be implemented by national standards (e.g., HIV/AIDS prevention, management of the blood system, the "Four Frees and One Care" AIDS treatment program, immunization programs, tuberculosis prevention and treatment, "Elimination of Tetanus and Reduction of Maternal Mortality," etc. - All townships should have at least one township health center owned and managed by government; all villages should have at least one clinic; the infectious disease regulation requires that local governments should finance disease control and prevention institutes, etc. 10 - A minimum financing requirement of RMB50 per beneficiary for the NCMS in counties implementing the scheme, with RMB10 coming from households, RMB20 from local governments at the provincial level and below (RMB40 in the case of relatively better-off eastern provinces and counties that do not receive central government financing); and RMB20 from the central government (western, central, and parts of eastern China). Since service provision is locally financed, however, compliance with national standards is not assured and has varied across regions and over time. Until recently, the approach was mostly laissez faire and decentralized. Although the first Education Law (1986) called for all localities to strive toward providing nine years of basic education and made it compulsory for all age-appropriate children, it provided a long and flexible timeline for meeting the targets. For example, the poorest provinces in the West were expected to extend free Rural Compulsory Education to only 65 percent of their population by year 2000 (West 1997). In recent years, the central government has taken a more active role in promoting rural public services and imposing short timelines for compliance. Since the late 1990s and especially in the past two to three years, in accordance with the new emphasis on a more balanced development strategy, the central government has taken a more active role not only in promoting new standards but also in providing financing, as it has increased efforts to improve public service delivery and social protection, including in the rural areas. These can be seen in the new sectoral initiatives such as the free Rural Compulsory Education program, the New Rural Cooperative Medical Scheme (NCMS), the Rural Medical Assistance Scheme, and the rural minimum living stipend schemes (dibao). The programs for the Rural Compulsory Education and NCMS are outlined in Table 2-1. In line with the promotion of a Harmonious Society, it can be expected that the central government will continue its activist role and increase its contributions to local service provision. Table 2-1: Key New Programs for Rural Education and Health Program Launch Policy Objective Policy Content Date "Two- 2003 To reduce financial costs Government provides funding to replace exemptions of schooling to families for revenues from the textbook and miscellaneous and one nine years of compulsory fees and provides a subsidy to boarding subsidy" education in order to students from "poor" families. expand access. Free Rural 2006-2007 To take on nine years of Government provides funding to replace Compulsory compulsory education revenues from "miscellaneous fees" (zafei) at Education financed by public an average of RMB140 per student p.a. for all resources; to reduce rural primary school students, and RMB180 financial costs to families per student p.a. for all rural junior middle in order to expand access. school students. By 2007 this will cover 150 million students. 11 Program Launch Policy Objective Policy Content Date New Rural 2005 To provide risk-pooling Designed mainly for in-patient services; run at Cooperative for major illnesses, to the county level, scope of coverage and Medical reduce the financial risks reimbursement rates are stipulated; Scheme of farmers falling into participation is voluntary and on a household poverty due to illness. level; counties are permitted to set up a NCMS when 70 percent of households agree to participate; cost-sharing: Center RMB20; subnational governments RMB20-40 per participant toward the annual premium; RMB10 by the participant. Source: Authors. 2.2. Financing Public Services Although the central government is not directly involved in the financing of most rural public services, it can powerfully influence the flow of resources. The central government's primary instrument is the use of legislation or regulations to mandate local government spending on services. At the same time, the central government indirectly supports rural services through revenue assignments and transfers to local governments. One example of the use of legislation to mandate local spending is the new Law on Compulsory Education (2006),19 which states in Article 42 that: "The government will put compulsory education into protected areas of public finance, and recurrent funding for compulsory education will be guaranteed by the State Council and local governments at various levels in accordance with this law." This amendment to the compulsory education law followed by five years the State Council directive, "Decision on the Reform and Development of Basic Education" (July 2001), which declared that the costs of rural compulsory education should be borne primarily by government rather than farmers and called for counties to take over responsibility for the salary payments of rural teachers, removing it from the township level. The ability of local governments to comply with these central government directives depends on their available fiscal resources. One of the sources available to local governments is own taxes. In China, since all local taxes are assigned by the central government and local governments have no discretion in setting either tax rates or tax bases, the availability of local revenues is largely determined by the central government.20 Since 1994, when the Tax Sharing System reform recentralized revenues, local revenues have fallen as a share of total revenues. This trend continued for the county and township levels even after 1998, in spite of increasing concerns about the rising insolvency at these levels. The trend can be seen in Table 2-2, where the combined county and township revenues fell to only 17 percent of total government revenues in 2004. However, since overall revenues have grown robustly as a share of GDP, rural own revenues have also increased slightly, from 2.5 to 2.9 percent of GDP. 19Law on Compulsory Education (amended by the NPC on 06/29/2006; implemented from 09/01/2006). 20Local governments can set the rates of some taxes within a narrow band. 12 Table 2-2: Trends in Rural (County and Township) Revenues 1998 2000 2002 2004 Percent of total revenues 20.3 19.7 17.1 17.2 Percent of GDP 2.5 2.7 2.7 2.9 Source: Adapted from World Bank (2007a). While assigning a declining portion of total revenues to local governments, the central government is providing a majority share of its own revenues as transfers, and transfers have grown rapidly as a source of local finance.21 Intergovernmental transfers are used around the world to give local governments additional resources, for the purpose of filling fiscal gaps as well as specific gaps in service provision.22 Since the late 1990s, the central government in China has greatly increased intergovernmental transfers to address the problems of the large regional fiscal disparities and severe shortfalls in some critical public services. Today, transfers are just over 60 percent of central revenues, a share that is one of the highest in the world. Central transfers are mainly earmarked. This is shown in Figure 2-1 where earmarked transfers comprise the majority. This is one consequence of the poor information that the central government has about local governments (see next section) ­ an excessive reliance on earmarking to control moral hazard problems. Since 1998, many new "Fiscal Capacity Transfers" have been introduced and the trend has shifted toward increasing the share of fiscal capacity transfers.23 In fact, it will be shown below that many Fiscal Capacity Transfers have "earmarked" characteristics. 21 In this report, we follow adoption of the new definition of transfers in the past few years by the MOF that excludes tax rebates. 22To reduce fiscal disparities across regions, countries such as Germany, Australia, and Canada provide fiscal capacity equalization transfers. To compensate for benefit spillovers and to induce local governments to spend more on services that are considered of high national priority, countries often offer specific purpose, matching transfers. To help local governments meet national minimum standards in selected services ­ and education and roads are among the most common ­ there are often specific purpose grants. 23Until approximately 2003, transfers were presented as "subsidies to local government expenditures" in the annual budget speeches by the Minister of Finance to the NPC and in the CFY. Since then, the MOF has begun to discuss transfers in a more differentiated manner, presenting a new category of "fiscal capacity transfers." Some figures for transfers used in speeches and publications exclude tax rebates, but this is done inconsistently. The CSY continues to report "subsidies to local government expenditures" as a gross figure that includes tax rebates. See World Bank (2007a, 2007b) for a more detailed discussion. 13 Figure 2-1: Trends in the Breakdown of Central Transfers 80% 70% Earmarked Transfers 60% s erfs anrTlatoTfo 50% 40% e 30% arhS Fiscal Capacity Transfers 20% 10% 0% 4 199 95 96 97 98 0 1 2 3 04 19 19 19 19 1999 200 200 200 200 20 Source: Calculated from CFY; Li and Xu (2006). The trend has been toward an increasing number of earmarked transfer programs, many of which are for the provision of public services in rural areas. Most of these fiscal resources are allocated for rural infrastructure, for which central and provincial earmarked transfers now make up the bulk of government investments (see World Bank 2007a). However, the last few years have also seen a proliferation of earmarked transfers for social services in rural areas. For example, by one count, there were 22 programs of earmarked funding for compulsory education in 2003 (see Annex 3).24 A number of these programs overlap with each other. Despite the large number of transfers for social services, they add up to surprisingly modest sums. All of the grants for education provided a total of only RMB4 billion in 2003, just 1 percent of budgetary expenditures on education or 0.65 percent of total education spending (Table 2-3). Table 2-3: The Share of Earmarked Grants to Finance Education 1994 1995 1996 1997 2003 Total earmarked grants (RMB bn) 0.252 0.497 0.760 1.163 4.062 Earmarked grants as share of budgetary 0.21 0.35 0.45 0.62 1.05 expenditure on education (%) Earmarked grants as share of total 0.17 0.26 0.34 0.46 0.65 expenditure on education (%) Total budgetary expenditure on 117.5 141.2 167.2 186.3 385.1 education (RMB bn) Source: Wang (2006) and MOF. Fiscal capacity transfers, while often tied, e.g., to salary payments, provide local governments with more spending discretion. The official Chinese definition of "fiscal capacity transfers" includes general transfers, minority region transfers, wage increase transfers, rural-fee-reform subsidies, "Three Rewards and One Subsidy," and "other 24Zafei transfers have been added to this list. 14 fiscal capacity transfers."25 These totaled RMB381.2 billion in 2005. Figure 2-2 presents the breakdown of Fiscal Capacity Transfers into its major components. The largest transfers are the wage increase transfer, the general transfer, and the rural-fee-reform transfer. Figure 2-2: Composition of Fiscal Capacity Transfers in China in 1998-2005 400 350 5 300 B 250 RM sn 200 4 Billio 150 2 6 100 3 50 1 0 1998 1999 2000 2001 2002 2003 2004 2005 1=General transfers, 2=Grants for minority regions, 3=Wage increase transfer, 4=RFR transfer, 5= three rewards and one subsidy and other reward programs, 6=Other fiscal capacity transfers Source: Calculated from Li and Xu (2006). Transfers are equalizing ­ both for earmarked transfers and Fiscal Capacity Transfers. Given the large and growing income disparities across China's many regions, equalization is a critical part of any attempt to improve the provision of rural public services, especially since it is widely recognized that since the 1990s, inadequate fiscal resources have been a major cause of the shortfall in services provided by local governments. Data presented by the MOF appear to show that both fiscal capacity transfers and earmarked transfers in 2005 were distributed across provinces in an equalizing manner ­ that is, on a per capita basis more funding went to provinces with lower per capita GDP than to provinces with higher per capita GDP. These are shown in Figure 2-3, where the fitted lines have a downward slope for both types of transfers.26 This is a dramatic turnaround from the pattern in the late 1990s (World Bank 2002) and a remarkable and praiseworthy achievement. 25In a 2004 publication introducing the concept, MOF officials explain that "Fiscal capacity transfers refer to those whose funds can be allocated by the recipient governments on a discretionary basis as part of own revenues" (MOF 2004). 26Linear regressions performed on Fiscal Capacity Transfers, earmarked transfers and the sum of these two transfers all confirm the equalizing effect, with negative coefficients on per capita GDP of -0.014 (R-sq = 0.289), -0.008 (R-sq=0.330), and -0.022 (R-sq=0.203), respectively. There were 30 observations, excluding Tibet. 15 Figure 2-3: Per Capita Transfers by Province in 2005 1400 1200 Fiscal Capacity transfers Earmarked transfers 1000 Linear (Earmarked transfers) s erfs Linear (Fiscal Capacity transfers) 800 anrt ati 600 caprep 400 200 0 0 10,000 20,000 30,000 40,000 50,000 60,000 per cap GDP Source: Calculated from Li and Xu (2006). The results of central government directives plus increased transfers have been a dramatic increase in budgetary contributions to the financing of rural public services. For rural compulsory education, total budgetary appropriations rose from RMB32.5 billion to RMB132.6 billion during 1998-2004, raising their share of total education financing from a low of 53 percent in the mid-1990s to more than 80 percent in 2004 (Table 2-4). Total budgetary appropriations for all education rose from RMB134 billion in 1998 to nearly RMB400 billion in 2005. Almost all of the budgetary increase was from local budgets. Of the increase of RMB263 billion, only 12 billion came from the central government, with the rest coming from subnational budgets (see Annex 3, Table 2). Table 2-4: Expenditure Trends in Rural Compulsory Education 1995 1998 2001 2004 Budgetary Expenditure (RMB bn) 32.5 46.8 78.6 132.6 Total Expenditure (RMB bn) 61.1 81.2 110.2 164.5 Budgetary Share of Total Expenditure 53.2 57.6 71.3 80.6 (%). Note: Figures are somewhat different from those published in yearbooks (see Box 2-1). Source: Zhou (2006), corrected. However, as we will show in Chapter 3, there clearly are limitations to mandating local governments to increase spending and trying to fill gaps within service sectors through specific programs and earmarked transfers. For example, despite the great emphasis on improving education provision in the two most recent Five-Year Plans, the share of total expenditures devoted to education fell nearly every year during 1998-2005 in the central government's own budget, from nearly 4 percent to less than 3 percent. In fact, the overall share of budgetary spending also declined slightly during this period, pointing to the limits of central mandates. Moreover, despite significant efforts to improve the equalization of transfers, the fiscal abilities of counties and townships to provide public services vary tremendously across the country. 16 2.3. Central-level Monitoring and Evaluation A hierarchical system of management is critically dependent on strong mechanisms for monitoring and evaluation that can provide timely and accurate information to policy-makers on policy implementation and its impact, to alert them to the need to adjust policies, and to hold local governments accountable for results. Under the planned economy, China built extensive data reporting systems in all sectors. To a large degree, these systems continue to operate, albeit with significant problems. In the health sector, for example, the Ministry of Health operates an elaborate information system that reports data on health system inputs (financial, human resources, facilities, etc.), service outputs, and health outcomes on an annual basis. The system is based on a mix of administrative data collected at the facility level, periodic surveys, and surveillance (e.g., infant and under-5 mortality). The Ministry of Education also has a system of detailed reporting that generates data to fill at least two large yearbooks ­ the Education Statistical Yearbook of China and the China: Educational Finance Statistical Yearbook. Although the system generates rich information, it has some major shortcomings. Monitoring focuses on fiduciary compliance, whereas aspects of the process, quality, and impact of monitoring are often much weaker. As in other formerly planned economies, the reporting system in China was designed to track fiduciary compliance and focuses on financial information and inputs and outputs. Other aspects of service delivery, while improving, are still much weaker. Additional information is also collected. For example, the health sector collects elaborate data on birth and mortality rates, life expectancy, immunization rates, attended deliveries, nutritional status, etc., but this information, although improving, is also still much weaker. The system is highly fragmented, with limited "horizontal" sharing of information. The usefulness of reported information is hampered by the fact that data are collected in a fragmented way by different agencies to meet their own purposes and there is little sharing or reconciliation of information. For example, the MOF and MOE routinely report differing figures for education finance (see Annex 3). Data reporting is always "upwards" and there is limited "horizontal" sharing, i.e., across line agencies. Most significantly, the National Bureau of Statistics (NBS) is ranked as a Deputy Ministry and therefore has no authority to lead in monitoring efforts, including those for public service delivery. The result is a less efficient and less "independent" fragmented system. Data are often unreliable. This is partly related to the highly fragmented system in which data are not shared. Partly, it is also due to weak quality control and because of strong incentives at facility or local government levels to report specific levels or trends in some indicators. In education, for example, the data on enrollments are considered so unreliable that great efforts are made to collect additional information through separate channels when counties undergo the certification process for "achieving" universal compulsory education. Furthermore, there are significant discrepancies between locally reported information and the national aggregates reported by the central ministries and published in yearbooks. Table 2-5 shows, using the example of fiscal information, how 17 large the differences are ­ in 2004 the sum of locally reported subnational expenditures and central expenditures totaled RMB468.6 billion in budgetary expenditures on education, a figure that is RMB83 billion, or 22 percent, greater than the figure reported in the national yearbooks. These discrepancies in reporting imply either that the national system tracks local expenditures incompletely, or that local reports (to the Ministry of Finance) are considered unreliable and are routinely discounted by the central ministries ­ perhaps to eliminate double-counting. Since both sets of data are in the public domain in the form of official publications, however, this could be a source of some confusion for analysts. Table 2-5: Differing Reports of Budgetary Expenditure on Education RMB Total Recurrent Sum of Central Implied total bn government budgetary provincial government budgetary spending expenditure* expenditure (1) expenditure (2) spending (1+2) 1998 192.6 133.8 203.3 12.3 215.6 1999 192.7 152.3 228.7 12.7 241.4 2000 218.0 176.5 256.3 14.1 270.3 2001 263.7 210.8 305.7 17.2 322.9 2002 310.6 264.5 349.1 21.0 370.2 2003 335.1 293.7 385.1 24.0 409.1 2004 385.1 336.6 446.6 22.0 468.6 * Excluding earmarked expenditures for the urban education surcharge. Sources: Calculated from CSY and CFY. Finally ­ though it is hard to distinguish between cause and effect ­ data are not used systematically by the central (or provincial) government to monitor performance, to benchmark local governments against each other, or to check urban-rural differences. On the one hand, data are often not easily accessible, unreliable, and reported in categories that do not lend themselves well to analytical uses. On the other, because of limited demand there is a critical shortage of analytical skills, i.e., the "evaluation" part of M&E. This holds true for individual ministries; it also reflects the lack of a leading evaluation agency that undertakes or commissions evaluations of government interventions for service provision in order to make budget or other decisions. Few central government policies or programs are piloted or even nationally implemented with a proper baseline established, a rigorous impact monitoring conducted, or evaluations of the intervention routinely commissioned. There are hopeful signs of change such as piloting of performance or value-for- money audits. Routine financial audits are conducted after the fiscal year and checking is done for compliance with budget warrants. Recently, however, the National Audit Office has conducted sample audits that include some performance assessments. Moreover, data reporting is often improved when the government responds to real or perceived crises. Following the SARS (Severe Acute Respiratory Syndrome) outbreak, for example, the health reporting system was complemented by a real-time Web-based disease reporting system that reaches down to the township level and monitors and tracks disease outbreaks nationwide. 18 There are also increased efforts to use information in planning and to hold local governments accountable for performance. In the case of public health programs, the Ministry of Health develops annual work plans that are disseminated to provinces and municipalities at an annual National Health Conference and that are further disseminated to local governments through local work planning meetings. The Ministry of Health has also issued standards for the organization and activities of the CDCs including staffing standards, organizational structures, and standard operating protocols that specify the tasks, monitoring arrangements, division of labor, etc. Similar, albeit less detailed, standards have been issued for other institutions with public health responsibilities. These standards form the basis for performance assessments by local Health Bureaus. There are also a number of program- or disease-specific performance management mechanisms. For example, the system of government official performance assessments, which determines the promotion prospects of local politicians and officials, includes indicators relating to family planning, epidemics, HIV/AIDS, and preparedness for Highly Pathogenic Avian Influenza. National priority programs also have program- specific monitoring, supervision, and training arrangements. In addition, similar program-specific monitoring takes place outside public health ­ e.g., in relation to the operation of the NCMS. Figure 2-4 presents the current system of control and accountability in China's public health system. Figure 2-4: Vertical Control and Accountability in China's Public Health System Provincial / central government General · MOH Annual work plan · Standard operational protocols · Staffing standards Program or disease specific · Government performance assessment (e.g. family planning and EPI included as assessment criteria) County government · Fiscal transfers and project financing · Supervision, training & TA General · Appointment of directors · Budgetary subsidies to institutions · Letters of Responsibility Program or disease specific · Letters of Responsibility · Supervision (TB, EPI, etc.) Public health institutions Source: Wang et al. (2006). Nevertheless, the monitoring and evaluation system of the central (and local) governments still has significant scope for improvement in terms of being able to provide timely and accurate information to policy-makers on policy implementation and its impact, to alert them to the need to adjust policies, and to hold local governments accountable for results. In terms of monitoring, China's statistics are becoming more reliable as modern survey methods are implemented and as the system is gradually moving toward a greater transparency of data within the government as well as to the public. On the other hand, as the government's tight control over the population loosens, 19 the accuracy of many statistics is being challenged.27 Most importantly for public service delivery, the system struggles to effectively monitor the impact of migration, as even population figures, e.g., for illegal migrants or unregistered children, are hard to come by. In terms of evaluation, China still largely lacks a systematic assessment of its policies and programs for public service delivery, in terms of process, impact, and cost-effectiveness. Finally, and closely interrelated to the above points on monitoring and evaluation, the government could strengthen its systematic use of monitoring and evaluation information for decision making. What this might require is that the State Council or one of the key (non line) agencies under it takes the lead in systematically commissioning evaluation studies for government programs and ensures that these evaluations are an integral part of the policy decision-making process, i.e., that they have a systematic impact on the delegation of responsibilities, specific programs, fiscal resource allocations, etc. The role of the central and provincial governments is to provide the overall policies and the framework for the provision of rural services. Throughout the 1990s, because of limited fiscal resources under central allocation, the central government accepted large regional disparities in services as inevitable in a country as large and as diverse as China. To reduce the burdens on the poor and to spread the benefits of growth, the central government has now adopted a much more activist role in calling for raising the floor on service provision so that citizens in even the poorest, most remote regions can have access to certain nationally designated minimum services. In this endeavor, the central government is constrained by the hierarchical administrative structure; by inadequate information about expenditure needs and fiscal gaps at the grassroots levels; and by weak mechanisms for monitoring, evaluation, and enforcement. To ensure implementation of policies, it is not enough to only provide resources, and ways must be found to ensure accountability for results ­ both the carrots and the sticks. Under the hierarchical system of administration, the use of these carrots and sticks in China is by necessity indirect ­ by remote control requiring a much greater reliance on the creation of appropriate incentives ­ to motivate both local governments and service providers and to ensure demand-responsiveness and participation by clients. 27See OECD (2005, Chapter 5), Gale (2002), or Holz (2004) for the challenges facing China's statistical system. 20 3. LOCAL (COUNTY AND TOWNSHIP) GOVERNMENTS Rural governments at the county and township levels have major responsibilities for the provision of rural services.28 As outlined in Chapter 2, the higher-level governments provide the mandates for service provision, along with the fiscal resources through revenue assignments and transfers, but it is the rural governments at the county and township levels that deliver the services. In this chapter we focus on the evolving roles of these two levels of government and discuss their capacities and incentives for implementing central policies. We also complete the discussion of the problems of the existing financing mechanisms: inadequate overall levels of funding, insufficiently equitable allocations, and limited capacity and accountability of local (i.e., county) governments to the central government. This is an interlocking set of problems: better resourcing (or reduced responsibilities) may be a necessary precondition to improving monitoring and accountability. Enhancing the accountability of county governments is central to improving rural services, because county governments have been given responsibility for the human development and poverty reduction outcomes that are key to building the new socialist countryside. 3.1. Assignment of Service Delivery The actual division of responsibility for the provision of rural services has been evolving and differs across provinces, given the absence of any central and formal assignments among local governments in China.29 Nevertheless, the county is clearly the key government level for the implementation of public service delivery in rural areas. Over the last few years, authorities and resources for service delivery have been increasingly concentrated at the county level. The pilot reforms to establish a direct link for channeling fiscal resources from the province to the county will further reinforce this trend. Even more significantly, townships today are increasingly under the management control of the county, with very limited discretion of their own. The yixian weizhu reform accompanying the Rural-Tax-Reform (RTR) moved the responsibility for basic education from the township level to the county level beginning in 2001. Treasury management reforms in many localities have put salary disbursements of township officials at the county level. "Verticalization" has been occurring in many functional departments that moved PSUs from township management upward to the county level. Finally, many localities ­ including Jiangxi ­ have implemented a reform to put "township funds under county management" (xiangcai xianguan) whereby the cash 28Although villages are not formal governmental bodies, they make significant contributions to the provision of services, such as primary education, social welfare and relief, and subsidies to military families and village militias. In practice, they have an ambiguous role in the sense that they perform both "government" and "community" functions. However, their legal role is to be a "self-governing" body for the villagers and we therefore treat them separately in Chapter 5. 29The broad patterns were formed during the 1960s and 1970s, when, under the development strategy of "walking on two legs," rural areas were expected to finance their own services, in order to conserve resources for the industrialization and modernization efforts. 21 management function for township activities has been moved to the county level. Together, these changes have hollowed out township coffers as resources are moved to the counties and counties have been put in charge of the bulk of rural services (and PSUs).30 In addition, since the RTR there has been a significant consolidation of townships and administrative villages ­ Xincheng Town (Jiangxi), for example, was created from the merger of four townships. This trend will likely accelerate under the strong incentives introduced by the "three rewards and one subsidy" program in 2005.31 Although these changes may help to reduce overhead, at least in the short run they may also disrupt services. For example, counties play a crucial role in the key services of health and education. In education, they are mainly responsible for planning, budgeting, and supervision of schools, hiring, evaluation, and promotion of headmasters and teachers, undertaking school infrastructure renovations and construction, and monitoring whether the target of universal enrollment for compulsory education has been reached. Within these general responsibilities, practices vary across counties, prefectures, and provinces as to how they delegate responsibilities to lower-level governments, such as township governments. Similarly, the county is primarily responsible for health spending for rural areas. However, since budgetary accounts lump together public health expenditures with subsidies and health care spending for civil servants, it is difficult to interpret trends. The health budget for a central China county, for example, shows that expenditures for civil service medical care absorbed 44 percent of the entire health budget for the county and two-thirds of the discretionary project funds after personnel expenditures. Little goes to the townships and below (fieldwork information, November 2006). The responsibility of county governments to provide rural public services has not only expanded due to the transfer of responsibilities from municipalities and, in particular, townships, but it has also grown in scope as well. As mentioned in Chapter 2, in the past few years the central government has significantly stepped up the introduction of new programs to benefit rural residents. For rural governments, these recent policies and standards impose harder targets and much shorter timelines than the policies in the past. Moreover, these policies are being imposed in an environment in which local governments are operating with less flexibility in the allocation of resources. During earlier periods, the government gave the collectives wide latitude to manage service provision to accommodate their varying revenue capacities.32 The increased dependence 30This "hollowing out" of the townships was vividly illustrated by Xincheng Town, where the mayor counted RMB1.68 million in township expenditures that are now directly disbursed from the county treasury and explained that the amount he was able to allocate at his discretion for the town government was only RMB325,000, until a special earmarked fund was created this year by the central government to return some discretionary funding to township governments. This fund provided RMB100,000 per township. 31This program will give a one-time reward of RMB500,000 for eliminating one township and RMB4,000 for cutting one staff position from the 2004 level. All "fiscally poor" counties are eligible for this program. Details are offered in World Bank (2007a). 32 In education, the collectives built minban schools ­ literally "schools run by the people" ­ with whatever resources they could muster and staffed them with minban teachers who were paid at levels similar to commune members. Production teams and brigades set up health clinics that were staffed with "barefoot 22 on "tied" transfers, the efforts to reduce farmers' contributions for public goods and services as well as those to bring extra-budgetary resources into the budget, and more stringent standards and specifications on how to implement public service provision all restrain county governments in their ability to make decisions according to local conditions. Additional capacity is also required for the implementation of other related policies for rural development, such as requests to increase village democracy and to participate in public service delivery. The new policies and the speed of their introduction impose immense pressures on local officials. In many cases, the policies have been implemented without sufficient guidance and capacity-building activities organized by higher-level governments. 3.2. Financing at the Local Government Level Own Revenues In contrast to the high and increasing expenditure assignments, the share of county and townships in own budget revenues is low and falling. The contrasting trends are shown in Table 3-1, where it is seen that since the 1994 Tax Sharing Reform, the shares of total revenues accruing to rural governments at the county and township levels have declined steeply, from 30 percent to 20 percent by 1998, and they continued to fall thereafter. The declining revenue share put downward pressure on rural expenditure shares, which fell from 31 percent of the total in 1992 to 26 percent in 2000. The expenditure trend was reversed when the central government began raising civil service wages in 1998, and these upward pressures on expenditures were exacerbated by central policies calling for increased services and service standards. Table 3-1 shows the vertical fiscal gap in the rural sector, growing from 8 percent of the national budget in 1998 to more than 14 percent in 2004. In GDP terms it grew to 2.8 percent. Table 3-1: Trends in Shares of Rural Public Finance 1992 1998 2000 2002 2004 Rural share of total revenues 30.1 20.3 19.7 17.1 17.2 Rural share of total expenditures 30.9 28.2 26.2 28.6 31.3 County 21.6 19.9 18.9 21.9 25.2 Township 9.3 8.3 7.3 6.8 6.1 Fiscal gap for rural sector* -0.8 -7.9 -6.6 -11.6 -14.2 As % GDP -0.4 -1.3 -1.6 -2.6 -2.8 * Calculated as the difference between revenue shares and expenditure shares. Source: Adapted from World Bank (2007a). In response to the increasingly evident fiscal problems of rural governments, the central government has introduced two new programs to boost rural revenues, in particular the reform to "complete" the Tax Sharing System reform introduced in 2003 doctors" who worked part-time in the fields and were paid from the net income of the collectives. After adopting the 1980s central mandates to staff their PSUs with state employees, local governments now have much less financial flexibility. With the Tax Sharing System concentrating revenues at the central level, and the RFR having eliminated a substantial part of their revenue base, many rural governments are operating with far fewer resources, and as a result they require greater assistance from transfers from higher-level governments. 23 and the "three rewards and one subsidy" introduced in spring 2005. Both measures rely heavily on the provinces to redistribute fiscal resources toward counties and townships, rather than any major redistribution of revenues from the central government. Completing the Tax Sharing Sytem reform. In December 2002, the State Council issued a "notification" to central agencies and provinces calling for provincial authorities to "complete the fiscal management system at subprovincial levels under the Tax Sharing System."33 The document calls on provinces to improve rural finance by rationalizing expenditure responsibilities among subprovincial levels, eliminating the practice of unfunded mandates, rationalizing the division of revenues, and improving the system of subprovincial transfers. More importantly, the reform explicitly holds provinces responsible for supporting rural public finance and improving equalization (World Bank 2007a). Three rewards and one subsidy. The second measure, the "three rewards and one subsidy" program, is designed to help restore a fiscal balance for rural governments by boosting incentives for revenue mobilization and the more painful downsizing reforms, as well as for provincial equalization efforts. The incentive program has two parts. The first part is a grant to county governments that have either improved their fiscal balance through above-average revenue mobilization or have reduced organizations and staff size. The second part is a grant to provinces that increases their equalization transfers to counties. There will also be an incentive program to encourage provinces to build a monitoring and accountability framework to "regulate and monitor" the fiscal performance of local governments. Although these efforts are laudable, the program is very small in size, with an injection of only RMB15 billion in central government funds in 2005. The beneficial effects of these two measures likely will not be felt until a few years after their introduction. In the meantime, though, other reforms were having exactly the opposite effect on rural revenues. Chief among these were the Rural Fee Reform (beginning in 2001) and the abolition of the agricultural taxes (2003-2006), which substantially reduced revenue bases at the county and township levels.34 The impact is felt especially in poor agricultural regions, where other tax sources are scarce (World Bank 2007a). The earlier reassignment of enterprise and personal incomes in 2001 from local to shared taxes had a similar impact in reducing overall rural revenues. The elimination of agricultural taxes (including the relatively rapidly growing agricultural special products tax) has especially limited options for rural governments to increase revenues, further reinforcing their incentives for infrastructural and other spending that attracts investments in industry and commerce, activities that compete directly with services for the ever-scarcer fiscal resources. Table 3-1 above confirms that, at least through 2004, aggregate data show a trend of continuing assignment of fiscal resources away from, rather than toward, rural governments. International experience strongly points toward the inefficiency stemming from a large fiscal gap between local governments' own revenues and their expenditure assignments (see also Box 3-1). Perhaps the most widely discussed and experienced 33State Council Document No. 26, December 26, 2002. 34This does not, of course, diminish the overall very positive impact that the reforms had on farmers. 24 problem is that of the fiscal profligacy of local governments, which over-spend in expectation of greater transfers or "bail-outs" by the center or province. For example, cross-country evidence on this is provided by Rodden (2002), showing that the dependence of subnational governments on transfers is positively associated with higher subnational fiscal deficits. The debt of counties and townships, including that of their service providers, is clearly an issue for China.35 Several countries have adopted formal deficit and debt rules, expenditure limits, and transparency mechanisms to try to contain such perverse incentives at local levels. An important additional financial resource for counties and townships is extra- and off-budget revenues. However, the rural-fee-reform has greatly reduced the scope for fee revenues and cut their contribution to service provision. While this policy to reduce farmer burdens is greatly welcome, it has had a negative impact on financial resources available for public service delivery that goes beyond the effect on revenue shares discussed above. Transfers The dependency of county and township governments on transfers is large and growing. Nationwide, counties on average are dependent on transfers for two-thirds of their expenditures when tax rebates are included (World Bank 2007b). In Dayu County, a middle-income county in Jiangxi Province, transfers (net of tax rebates) have risen steeply from a level of only 9 percent in 1998 to account for nearly half of all budgetary expenditures at the county and township levels in 2004 (Figure 3-1). However, this mainly reflects the declining portion of own revenues and the growing expenditure responsibilities of these local governments. 35Information on local debt is relatively weak. However, a study undertaken by Mei Wang (see also World Bank 2007a) of 57 counties in three provinces and one municipality suggests that county-level debt can be as high as 14 percent of GDP. The average ratios of debt to own, disposable, and total revenue are estimated to be, respectively, 335 percent, 255 percent, and 138 percent under a best case scenario; and 558 percent, 425 percent, and 231 percent under a worst case scenario. Even the lowest debt to own revenue ratio (138 percent) is much higher than the international standard for developing countries, which ranges from 60 percent (as measured by Standard & Poor's), or the level of some individual countries such as Colombia (80 percent). 25 Figure 3-1: Rising Dependency on Transfers in Dayu County 60% 50% 40% 30% 20% 10% 0% 1998 1999 2000 2001 2002 2003 2004 Source: MOF county-city data, various years. Most transfers are just co-financing added costs or replacing lost revenues due to policy changes and are not adding revenues for county and township governments. For example, among the largest transfers introduced in recent years is the wage increase transfer, which was RMB148 billion in 2005, accounting for 21 percent of total transfers or 39 percent of "fiscal capacity transfers." This transfer is designed to defray the added costs of civil service pay increases mandated by the central government over the past few years and it is allocated to local governments according to their post establishment requirements. However, the transfer does not cover all the local government costs of these pay increases, even in the poorest localities.36 The rural-fee-reform (RFR) transfer, another major transfer, is also not replacing the full costs of the policy, i.e., local revenue losses. The rural-fee-reform transfer, which includes compensation for the agricultural tax reductions, is intended to replace revenue losses of local governments due to the elimination of fees and agricultural taxes. However, the RFR transfer, which was RMB66 billion in 2005, accounting for 9 percent of total transfers and 17 percent of the Fiscal Capacity Transfers, leaves a gap estimated at one-third to two-thirds of previous revenue levels (World Bank 2007a). Revenue replacement is the main objective of the central earmarked transfer for free Rural Compulsory Education. Under its predecessor program of TEOS, the central government provided a textbook subsidy to students from poor families to offset the textbook fee that would be exempted for those students ­ one exemption. In the current program, the central government provides a per student transfer in exchange for schools stopping collection of all miscellaneous fees ­ that is, replacing an extra- budgetary revenue with a budgetary injection from the central government. In this as well, the standards appear to be too low, with schools complaining that at the current levels, central subsidies are providing only a small fraction of what the schools 36None of the localities visited report receiving full coverage from this transfer. The reasons vary. Sometimes it is by design: coastal and central areas are expected to cover part of the costs from own resources. Sometimes it is because staffing is in excess of norms (post establishment), and the subsidies do not cover the excess, etc. 26 previously collected in various fees. 37 Since the program intends to make Rural Compulsory Education truly "free of charge" by prohibiting the collection of any fees, setting the replacement subsidy at a level that is too low will lead to a net loss of revenue in many schools that could become severe. Among all three transfer types, because of the low funding levels, none of them is revenue- or cost-neutral, far less than making any net contribution toward enhancing the capacity of rural governments to deliver services. Consequently, in most localities it appears that, despite growing transfers, fiscal resources are not sufficient to ease the financial strains imposed on rural governments, since they are not able to keep pace with rising expenditure needs driven by central policies. Overall, because of inadequate compensation of revenue losses, transfers are doing little to address the problem of inadequate revenues or insufficient incentives in the rural sector to provide public services at mandated levels. Instead, these transfers are associated with new programs that impose additional expenditure requirements on rural governments. In some cases, the low standards for revenue-replacement transfers may result in excessive costs being pushed down to local governments. For example, in Chongqing Municipality (a western regional unit), the TEOS requires 20 percent co-financing from subnational governments. Chongqing has decided to cover the entire local portion for the eighteen poverty counties. In the other county-level jurisdictions, the local portion is to be shared equally with the counties or districts, so that the total subsidy is shared, i.e., 80 percent central, 10 percent municipal, and 10 percent county. The subsidy standard is set at RMB170 p.a. for primary students and RMB250 for junior middle students in rural schools. The subsidies for county and town schools are RMB200 and RMB280, respectively. On average, program costs are perhaps RMB20 per enrolled student in basic education for governments in non-poor counties, 10 percent of the subsidy to the schools. The actual costs to the counties may be much higher, because schools were collecting far more in fees prior to 2006 ­ one middle school in Guangxin Town reported that it was collecting RMB1160-1360 p.a. in miscellaneous and activities fees until it was ordered to stop in 2006. The county governments may have to step up their own budget appropriations to the schools to make up for part of the shortfall (of upward of RMB1,000 per student!) to avoid further increases in indebtedness and service reductions and to prevent the "missing revenues" from being passed on again to the students in new guises. Local governments are also invariably expected to absorb the administration costs for new programs. This could impose considerable costs given the large numbers of beneficiaries ­ as for the TEOS, NCMS, and social protection programs. Box 3-1: Perverse incentives of fiscal transfers In addition to a general problem of local accountability in converting resources into actual services, the institutions of intergovernmental sharing of resources and responsibilities themselves have implications for local accountability. In many developing countries that are 37Estimates from a sample of more than 3,000 villages show that the costs of schooling often exceed 20 or even 40 days of an average male daily wage, a sum far higher than the fees that are now being exempted (World Bank 2007d). In two middle schools visited in Chongqing, schools reported previously collecting close to RMB1,000 per year from students, a level more than five times as high as the zafei subsidy. 27 newly decentralizing, and where local jurisdictions have little own-revenue potential, over- dependence on national transfers undermines the accountability of local governments to their citizens. Perhaps the most widely discussed and experienced problem with over-dependence on transfers is that of the fiscal profligacy of local governments, which over-spend in expectation of greater transfers or "bail-outs" by the center or province. A second type of problem is that of misallocation of resources because of perverse incentives created by transfer design. Transfers for the salaries of public employees can lead to a bloated civil service or local administration, with public jobs being used by local officials to extend patronage, at the expense of quality service delivery through strong incentives of service providers and spending on other service delivery inputs. Transfers for capital asset construction discourage spending on recurrent maintenance and hinder optimal local-level planning for the prioritization and placement of assets. The incentives for public agents is to try to access more and more of these central transfers for new asset construction, while no agent claims responsibility and ownership to maintain the assets thus created. A review of international experiences suggests that developing countries have rarely been able to use conditional transfers to ensure minimum standards of service delivery. At best, the evidence suggests that conditional transfers have been successful in increasing spending by subnational entities in the areas that were included as "conditions" for the transfers. However, the more recent accountability concern is that increasing spending has not sufficiently translated into better services that work for the poor and disadvantaged. Generally, it is either because legal instruments to force states ­ in federal countries ­ to do things they do not want to do are weak, even in developed countries (as in the U.S. or Switzerland), or because central government officials rely on the support of local governments to remain in office (in Argentina or Brazil). With these concerns in mind, several countries are adopting more serious policy strategies to measure performance of services in reaching and benefiting the poor. Some, like Chile and Brazil, are pursuing transfers conditional upon service delivery performance. Others, like South Africa, are beginning to set up systems to measure service access by poor households and to use that as the basis for targeting resources to specific jurisdictions. Still others, like India, are beginning to set up systems to comprehensively monitor performance in service delivery by local governments across sectors, as a planning and accountability device to facilitate local decision making over resource allocations, but without conditioning transfers on these performance indicators. Source: Khemani (2006). The equalization effect of transfers is limited. Chapter 2.2 shows the improving, but nevertheless limited equalization impact ­ measured as transfer per person by province. A comparison of total per capita fiscal expenditures across provinces ­ financed from own revenues plus transfers ­ shows that interprovincial disparities grew during the 1990s and remained very large through 2004 (see World Bank 2007a, Chapter 4). This partly reflects the coarse measures for equalization that are applied when compensating localities for revenue impacts from policies including for public services. There is a tendency for flat rates to be applied. For example, under the TEOS, a flat rate of RMB70 is applied for textbooks for primary school, and RMB140 for junior middle school, for all eligible students. The central government ceilings for the TEOS transfers are presented in Table 3-2. The program is designed to "tilt" toward the poor regions, and how much each province receives under this program depends on its 28 location: 80 percent for western provinces, 60 percent for central provinces, and zero for coastal provinces.38 Table 3-2: Central Government Ceilings for TEOS Norms Primary Junior Middle (RMB) (RMB) Exemption of Miscellaneous Fees (zafei) 130 180 Exemption of Textbook Fees 70 140 Boarding Subsidy 200 200 Maximum of total TEOS an eligible student 400 520 can receive per year Source: Zhou (2006). The methodology applied to the distribution of the RFR transfer and the wage increase transfer is roughly similar. The western region ­ comprising twelve provinces with 28 percent of the nation's population39 -- receives transfers that cover 80, or 100 percent of the program costs. The central region, with ten provinces and 41 percent of the nation's population, receives the majority of the program costs from central transfers. The coastal region, with nine provinces and 31 percent of the total population, receives few or no transfers. Since even within each region incomes and expenditure needs are very diverse, a formula that treats these large areas and populations as uniform can yield only very coarse equalization. Paradoxically, in fact, in the 2006 version of free Rural Compulsory Education, where the zafei subsidy is based on what the province set as the zafei standard in the previous year, the outcomes may be counter-equalizing, for two reasons: · First, the way of calculating compensation for exemption of miscellaneous fees across provinces is inherently unfavorable to the poor provinces. The formula for allocation is based on enrollment figures reported the previous year, times the per student miscellaneous fees set by the province in the previous year. Because the fee level set by provincial governments is usually based on the ability of their own residents to bear the cost, poor counties and poor provinces set lower fees, whereas richer localities set higher fees. A formula based on the original fee level will transfer more resources to the richer provinces. · Second, the per-student formula for allocations favors big schools over small schools, which predominate in the poor, remote regions. A student-based allocation calculated to be sufficient to cover the operating costs of large schools will likely not be enough for small schools, that is, those with perhaps 100 or fewer students. Even the "equalizing" grants have limited effects on rural services because covering public sector salaries is the first priority in their allocation, rather than equalizing service provision standards per se. The "general transfer" and the minority region transfer are the equalization transfers that are targeted at poor and minority regions. They have grown rapidly in recent years, and this represents a significant improvement over 38In addition, free textbooks and exemptions for boarding fees are provided to eligible students in western areas and national poverty counties in the West, respectively. 39These population shares are based on the 2000 census. 29 the situation in the mid-late 1990s (World Bank 2002). However, the equalization effect of these transfers is diluted by the formulae currently in use, which define "expenditure needs" of local governments largely by personnel expenditures based on current staff numbers cum post establishment figures.40 As a result, although they contribute to reducing fiscal disparities across localities, these transfers do not provide strong support for the provision of rural services. In sum, in spite of the huge recent increases in resources, the transfer system is still not providing enough money for local governments to meet the growing expenditure responsibilities. For most localities, it appears that fiscal resources are not sufficient to ease the financial strains on rural governments, and transfers are doing little to address the problem of inadequate resources and insufficient incentives in the rural sector to provide public services at mandated levels. Finally, the intergovernmental fiscal system in China still lacks a mechanism for ensuring that a mandated service is actually financed. In the absence of such a mechanism, provision is still left ultimately to local finance. The result is shown in Figure 3-2, where the average per student expenditure in primary school in each province is very closely related to the province's per capita GDP. Figure 3-2: Primary School Spending by Province in 2004 8000 7000 total 6000 n) uay(tn 5000 4000 udetsr budgetary 3000 pe 2000 1000 0 0 10,000 20,000 30,000 40,000 50,000 60,000 per cap GDP (yuan) Source: Calculated from CEFY 2005; CSY 2005. 3.3. Compliance with National Rural Public Service Policies The difficulties of central policies not being fully implemented at the grassroots level has existed for a long time. When for several years the centralization of revenues left expenditure responsibilities for local government that far exceeded their revenue capacities, there emerged an acceptance that local governments could not be expected to fulfill all of their assigned tasks. This notion persists and is in itself a hindrance to any effort to enforce local accountability. Under funding remains a large obstacle to the enforcement of central policies in many counties and townships. For example, fiscal theory and international best 40Li and Xu (2006 p. 63). 30 practices both emphasize that allocating sufficient revenues is a precondition for effective grant programs. As seen in the last section, this condition is not met in most of the current grant programs in China, where program funding from central transfers is typically far below costs. In all major programs the central government requires that local governments share the costs, even the poorest provinces in the western region. Although the local share often appears reasonable for each program, it does not take into account the accumulation of costs from all programs being implemented. Moreover, central directives for rural service provision compete directly with local preferences for spending on infrastructure. In all localities there is a strong incentive to increase revenues, especially own revenues that local governments can allocate freely. Given the tax structure emphasizing industry and commerce, and especially since agricultural taxes have been abolished, seeking local revenue growth means that governments engage in growth-enhancing investments in infrastructure to attract businesses to the area. This incentive is reinforced by the convergence of interests in growth-promoting infrastructural investments for mobilizing both on-budget and off- budget revenues, as these investments raise the value of land locally ­ the main source of extra-budgetary revenues for local governments. To the extent that meeting central mandates on service provision compete with infrastructural investments for scarce fiscal resources, the strong conflict with local interests reduces incentives for compliance. As argued in the previous section, a heavy reliance on transfers creates perverse incentives for local governments. As shown in Table 3-1, despite the many reforms the revenue share of rural governments has continued to fall and the vertical fiscal gap has grown. Given that fiscal disparities across counties remain largely undiminished, many rural governments have inadequate revenues to meet their expenditure responsibilities for service provision (World Bank 2007a). The perverse incentive basically stems from soft budget constraints, especially for the payment of salaries. Consequently, local officials have little incentive to resist pressures to hire under a budgeting system where transfers are driven by "expenditure needs" that are calculated by the size of the fiscal-supported population (). Under the "three guarantees" rule of budget execution, guaranteed wage payments, guaranteed normal functioning of government, and guaranteed social stability, money can always be found for salary payments, whereas service provision inevitably takes a lower priority. The result has been rapid employment growth at the subnational levels of government: the fiscal-supported population grew from 35.15 million at year-end 1999 to 42.94 million in 2005, during a period when the central government underwent a drastic downsizing campaign that cut its own staffing levels by half. Nevertheless, nationwide, total public sector employment grew from 2.9 percent to 3.3 percent of the total population, growing annually at 3.7 percent (Li and Xu 2006, p. 64). Until M&E mechanisms are strengthened, there is a risk that as the central government increases transfers for rural services, these transfers are only replacing local expenditures rather than adding resources, as local governments divert their own funds to other uses. This appears to be confirmed by the trend in education spending in Dayu County (Table 3-3): 31 Table 3-3: Trends in Education Spending in Dayu County Year Education expenditure per capita (RMB) Share of budget spent on education 2001 90 15.3% 2002 104 14.0% 2003 112 13.6% 2004 123 12.3% 2005 142 11.0% Source: MOF county data. Another challenge to compliance is the sheer number of reforms and programs to be implemented by local governments. Over the past few years, local officials have endured an almost dizzying array of administrative changes that include adjustments in the administrative hierarchy governing grassroots units, consolidation of townships and villages themselves, as well as reorganization of township governments. Many of these administrative changes have a significant impact on how policies and resources are transmitted to the rural sector. Reforms are sometimes running at cross-purposes: the RFR is reducing revenues at a time when other measures are increasing expenditures. They are shifting authorities across levels of government, stripping townships and villages of resources, and hence decreasing incentives for service provision. The instability created by all these changes undermines the planning and rationalization of resource allocation. Moreover, new programs in basically every sector ­ for social services ranging from TEOS and free Rural Compulsory Education in education to CDCs, NCMS, Medical Assistance, and other initiatives in health to wubaohu, dibao, and other reforms in social protection ­ require time as well as analytical and design capacity at the local level, and thereby stretch county and township capacity and ability to comply, at least in the short run. Moreover, the proliferation of transfers constrains local government budget autonomy and reduces transparency. The ad hoc funding levels and uncertain timing of its arrival further undermine any meaningful budget process at the local level. The proliferation of transfers also leads to reduced transparency, including at the local level, that leads to a lack of accountability and often prevents local governments from effectively coordinating various programs. Too rapid implementation of changes carries costs. Local governments have little incentive to show willingness/ability to pay for the new programs since over the past few years the central government increased transfers when it appeared that local governments could not pay. For example, the progression of the TEOS illustrates how the central government started out with the task of financing textbooks for children from poor families, but within three years expanded to paying for the zafei of all rural children. Although these central policy changes are all to the benefit of the rural populace (and thus, good), they do create expectations of ever-greater largesse from the central government and thus discourage a local response. Accountability is also difficult to enforce because of the fragmentation of authorities in this system, whereby higher levels of government have limited means to enforce established standards or targets or to influence the incentives of lower-level governments. The level-by-level management of the health system presented in Annex 1 shows that higher-level governments exercise little financial leverage through the transfer systems, 32 whereas health institutions at higher levels do not exercise any direct administrative or legal control over institutions at lower levels. For example, on issues such as the appointment of the director, staffing decisions, the budget process, and the setting of performance targets for a county-level CDC, it is the local Health Bureau rather than the city or provincial CDC that has formal responsibility. Accountability hence tends to be through promotions in the personnel or Party systems or through high profile punishment of transgressors.41 In the Chinese system there are many tools to hold officials personally accountable for outcomes ­ family planning targets, tax collection targets, etc. However, until the systemic constraints that impede individuals from fulfilling their functions are removed, personnel accountability will be hard to enforce. For example, the Chinese education system holds principals personally accountable for ensuring that schools' needs are met, but they have no control over the resources that are needed to support schools to fulfill all their functions; teachers are accountable for student performance but have no control over the levying of fees and charges that may result in some students dropping out from school activities. The widespread use of accountability measures that ignore systemic influences on outcomes renders them ineffective since they would be widely considered unfair and unreasonable and would tend to elicit either "rightful resistance" or, more commonly, collusion among those affected to falsify reporting as necessary to ensure that everyone meets targets. The rigid personnel system with life-long tenure also severely limits the use of personal accountability. The challenge is to adapt the system of personal responsibility to an environment of multiple objectives, including many public service delivery outcomes that are relatively difficult to measure. Although the personal responsibility system has enabled the central government to ensure compliance with some key priorities, the system tends to work best for programs with relatively simple and measurable objectives and outcomes. It is less likely to work well for programs with complex, multiple objectives, where outcomes are not easily measured and compared across localities. 41For example, Vice Premier Wu Yi has made repeated promises to punish any officials found concealing the extent of epidemics or failing to properly implement national treatment and health care plans. 33 4. SERVICE PROVIDERS In this chapter we examine the next critical layer between central policies and their implementation ­ service providers, such as schools, hospitals, clinics, nursing homes, etc. ­ and we identify the weaknesses in the accountability relationship between local governments and service providers, in particular the public service units (PSUs) that dominate most service sectors. These weaknesses stem from unclear assignment of responsibilities, the under funding of public services and lack of financial accountability of PSUs, and lax performance management, all of which are exacerbated by the lack of competition. Consequences are failures in rural public service delivery. 4.1. PSUs and Delegation by Local Governments The vast majority of public services in China are provided by PSUs, public entities that are attached to government at various levels. These include most schools, agro- technical extension stations, cultural centers, health clinics, and hospitals, etc. 42 Nationwide, there are over one million PSUs employing nearly 30 million staff, accounting for some three-quarters of public employment (excluding SOEs and the armed forces). Most PSUs operate in rural China. According to a recent study by the NDRC, over 80 percent of PSUs are affiliated with county and township governments, employing 65 percent of the national total of the PSU labor force (NDRC 2004). Education constitutes the largest part of PSUs and is the most tightly linked to government. Rural compulsory education is provided by some 400,000 regular schools and 100,000 "teaching points" to about 143 million rural students in grades 1-9. Almost all of these schools are PSUs, with the exception of a few village schools that are independently managed. In contrast, private providers play only a minor role to date, especially in basic education.43 In contrast to education, the health sector is characterized by a wide range of different providers and institutions. Inpatient care is provided by health centers and hospitals at township levels and above; the most important providers (in order of utilization rates) are: county hospitals, THCs, Traditional Chinese Medicine hospitals, and city hospitals. Outpatient care for rural residents is primarily provided at the village and township levels, although county hospitals, Traditional Chinese Medicine hospitals, and hospitals above the county level also play a role. Most likely, many patients self- treat through private pharmacies, but data on this are limited. The organization of public health programs and delivery of public health services are more complicated. Broadly speaking, the county Health Bureau has overarching responsibility for most areas of public health at the county level and exercises this responsibility through the CDC, MCH 42 The State Council (1998) decree on PSU registration defines a PSU as a "social service organization established by the state for the purpose of social public benefit." 43According to China Education Yearbook(2005), there were 6,242 private primary and 4,608 private junior middle schools in China, accounting for 6.1 percent and 6.0 percent respectively of total enrollment at that level. The role of private providers is significant and growing in terms of serving the children of migrant workers in the cities. However, the numbers of private providers may not be fully captured in the official statistics. 34 Center, and, in some localities, disease-specific institutes (e.g., tuberculosis).44 Some of these providers are PSUs (county hospitals, most THCs, CDCs, MCH Centers, etc.). Most village clinics and some providers at higher levels are private entities. Table 4-1: Overview of Health Care Providers Inpatient care Outpatient care Prevention and public health County County hospitals, private County hospitals, County Health Bureau, CDCs, MCH clinics, MCH Centers private clinics, Centers, Family Planning Stations, (deliveries) pharmacies Disease-specific institutes, Center for Health Inspection and Supervision, County hospitals Township THCs, private clinics THCs, private Township administration, THCs, (limited in cases of inpatient clinics, Family Planning Stations care) pharmacies Village Village clinics, Village clinics, Family planning cadres, pharmacies Informants Source: World Bank (2007c). Despite the many variations in PSUs and their relationship to government, there is no differentiated framework for PSUs, e.g., to make a formal distinction between the two categories of PSUs, delegated PSUs and devolved PSUs (Box 4-1). The rationale for such a distinction lies in the fact that the performance of most non-core executive functions requires professional technical skills rather than policy-making or administrative skills. The typical international example is the executive agencies of the UK government, which are separate from, but in the meantime affiliated with, core government departments. In the Chinese context, the separation of the Labor and Social Security Bureaus from the Social Insurance Management Authorities is an example. Such a separation may also be needed between Health Bureaus and the management authorities of the New Cooperative Medical Scheme.45 Box 4-1: Delegated and devolved PSUs A delegated PSU can be defined as one that is directly affiliated with a core government department without an independent legal status and accounting system, and governed by a manager appointed and supervised by the head of the government department. The manager has relatively limited autonomy in operations, which is larger than a core government department and smaller than a devolved PSU. Monitoring is more of the nature of inputs and process monitoring than results monitoring. The personnel system of a delegated PSU may be integrated with that of 44 Family planning stations fall under the responsibility of the State Family Planning Commission, which has local representation. In addition, some public health agencies operating at the county level are effectively local branches of central agencies (e.g., the State Food and Drug Administration, Environmental Protection Agency). But hospitals, health centers, and village clinics are also involved in public health. In the case of county hospitals, this primarily involves immunizations and MCH services (i.e., considerable overlap with MCH Centers). At township and village levels, dedicated public health agencies are not represented, so these providers become the main vehicles for providing both personal (immunizations, well baby care, tuberculosis screening and treatment, etc.) and collective (health education, sanitary inspections, environmental health, etc.) public health interventions. 45Once these two organizational forms are created through central government decree, decision-making rights can be left to provincial and county governments with respect to which is more applicable to existing PSUs. In general, however, public schools and hospitals may find devolved PSUs more appropriate, while PSUs that execute government administrative functions can be organized as delegated PSUs if there is a rationale for them not to be integrated with core government offices. 35 the core departments and put under the management of the Post and Establishment Offices and Ministry/Bureaus of Personnel at each level of the government. Revenue-generating activities of delegated PSUs should be minimized and put under the same degree of close scrutiny as core government departments. No delegated PSU should be allowed to borrow. A devolved PSU, on the other hand, may be defined as one having an independent legal status and accounting system. Its internal governance may consist of the following components: (i) Every devolved PSU should have its own mission statement and a list of performance indicators that can be evaluated periodically to determine the extent to which the objectives have been achieved. (ii) Some devolved PSUs may be governed by government department-appointed board that acts as their owners, or acts as policy-maker, with proper participation of consumers and representatives of the other stakeholders. (iii) Appointments and dismissals of managers may be determined by the board wherever it is created. Other hiring and firing decisions may be retained by the board (e.g., in the case of key technical and managerial staff) or be delegated to the management. The board and the management should be given autonomy to establish for their staff a human resource management system that they believe works better in their unit. (iv) Revenue generation of devolved PSUs can be allowed, but subject to strict rules of accounting, auditing, and reporting. (v) Borrowing by devolved PSUs should be either prohibited, or approved by the Bureaus of Finance from which they receive their budget. Source: World Bank (2005a). China's strong reliance on public service entities differs from that in many countries where the government purchases more public services from private providers or other non-public organizations, which might then operate on a for-profit or not-for-profit basis (World Bank 2004a, 2004b, 2004c).46 In many public service sectors in China, including education, hospitals, and agricultural services, until recently only PSUs were allowed to operate. However, reforms have been underway since 2002 to change the status of some PSUs to give them more operational autonomy and to open up services to provision by private and nonprofit entities (World Bank 2005a). Public finance for public services is channeled almost exclusively to PSUs. In China, the boundary between PSUs and other service providers is clear insofar as budgetary support goes almost entirely to PSUs. Even where private entities such as village health stations provide services that might be regarded as "public interventions," such as childhood immunizations and sanitation work, they receive little assistance from public resources. Consequently, the following analysis focuses on PSUs. Experience from pilots on the provision of public services through non-public entities is limited. In recent years China has begun to experiment with "buying services with money" (huaqian mai fuwu). Two such experiments are presented in Box 4-2, both illustrating that government may capture efficiency gains and offer consumers more control over services by separating the financing and service provision functions. These reforms are only in an embryonic stage, and bear careful watching. Similarly, a pilot in Hubei represents an invaluable experience of rural public service reform in China. As part of the "rural comprehensive reform," the pilot built on the achievements of the tax- 46For example, in the U.S., a study found that in a sample of sixteen communities in the early 1980s, only 39 percent of social services, employment and training services, housing and community development, health, and arts and culture were delivered directly by government departments, while 42 percent were delivered through nonprofit organizations, and 19 percent through for-profit organizations (Stanton 2003). 36 and-fee reform and fundamentally changed the institutional foundation of rural production and cultural services delivery. Specifically, the reform changed the role of government in service delivery, from a "general manager of providers" sitting on the supply side to a "general purchaser and evaluator" sitting together with consumers on the demand side. This change significantly strengthens the bargaining power of farmers. For more details see Annex 4. Box 4-2: Buying public health services in Chongqing and Wuxi On August 17, 2006, the People's Daily reported two stories about public health reform and expressed the editor's appreciation of their value to other regions. The first story involves a reform in Qianjiang District of Chongqing Municipality. Farmers constitute over 80 percent of the population of Qianjiang District. In July 2005, the district government launched a reform to strengthen public health services provided to rural women and children. Instead of subsidizing the Maternity Health Center of each township, as is the case almost everywhere in rural China, Qianjiang decided to use the money to subsidize the users of the services through a voucher program, whereby farmer families received vouchers for four sets of public health services, namely, vaccinations, health checks for pregnant women before giving birth, post-birth visits, and health checks for children. Holders of the vouchers were free to choose any Maternity Health Center in the district, not necessarily the center in their own township. The second story is about Wuxi City of Jiangsu Province, where there have been good results by contracting a private hospital to run the city's tuberculosis prevention program. Source: People's Daily, http://paper.people.com.cn/rmrb/html/2006-08/17/node_1.htm. Local governments have introduced performance compacts with PSUs and other service providers. For example, schools have been asked to sign "Education objective- accountability agreements" [mubiao zerenzhi] specifying the provision of education services, even though their design and significance vary tremendously across counties (Liu 2007). Similarly, most counties rely on Letters of Responsibility to set out priorities and performance targets for health institutions. These are generally signed between the Health Bureau and institutions such as the CDC, MCH Centers, county hospitals, and THCs.47 Nevertheless, these contracts explain only part of the relationship between the government and PSUs, and the boundary between government and many PSUs is often blurred. As the last part of the public sector to undergo reform, PSUs are the last vestiges of the planned economy where government spills over into PSUs and vice versa.48 Some PSUs have multiple functions, from executing administrative functions (e.g., animal quarantines) to providing commercial services (e.g., castration of pigs). PSUs are often asked to perform government tasks without compensation. Hospitals are 47Letters of Responsibility can also extend to village clinics. For example, in one of the counties in the county case study undertaken as part of the study on rural health (World Bank, 2007d), village clinics sign a Letter of Responsibility with the THC, which sets out indicators for reporting on infectious disease, MCH, and clinical services. The THC monitors the activities of the clinics, and the renewal of licenses is conditional on a satisfactory performance review against the indicators in the Letter of Responsibility. 48In the planned economy the public sector encompassed government, PSUs, and SOEs. SOEs have gradually been separated from government and since the 1990s have been delinked from the public purse. Core government departments have undergone several rounds of reorganization and restructuring. Until recently, what were called "reforms" in the PSU sector consisted mostly of liberalization steps that gave PSUs increasing autonomy to raise own revenues and to determine their own scope of business. 37 expected to provide care to indigent patients as well as to uninsured victims from traffic and other accidents, for which they often receive no compensation from public resources even though they are almost entirely self-financed. Some PSUs provide staff and other services to government departments "on loan," etc. More informally, individual PSU employees, including, for example, school teachers, are sometimes mobilized as "cadres" to execute administrative functions for the township. PSUs are also often asked to take onto their payrolls demobilized soldiers, university and technical school graduates, staff made redundant by downsizing in government departments, and even the children of government employees in supervisory departments, etc. This is a major cause of the "overstaffing"49 at PSUs as well as the preponderance of unqualified staff, especially in agricultural extension services and health institutions (World Bank 2002, 2005a). Management of PSUs is fragmented and left mostly to functional departments. Despite the substantial weight of PSUs in county and township budgets (see Section 4.2 below), no department in local government has comprehensive overview of PSUs to track what fiscal resources they are absorbing or how efficiently they are using them, etc. Instead, management of PSUs is scattered under different government departments ­ the planning department, the Finance Bureau, and virtually all the line bureaus to which the PSU reports (education, health, agriculture, etc.). Interviews with local government officials revealed that none could provide a comprehensive overview of the financial situation of the PSUs in their charge. 4.2. Financing of PSUs The financing of PSUs is a significant burden for local governments. Nationwide, PSUs account for 30 percent of national budgetary expenditures (World Bank 2005a). Their share is greatest at the county level, where the weight of social services, in particular, is heavy (Chapter 3). The different shares of PSUs in government budgets are shown in Table 4-2, where available data for 2002 show that 54 percent of county expenditures were spent by PSUs. For counties and townships together, 46 percent of the combined budgets was spent on PSUs. These shares have probably grown since 2002. Table 4-2: Shares of PSU Financing in Budgets at Different Levels in 2002 Budgetary Total budgetary Share of PSUs in expenditures on PSUs expenditures total Central Government 116.8 677.3 17.2% Provinces 154.0 433.1 35.6% Municipalities 100.0 463.7 21.6% Counties 261.5 482.0 54.2% Townships 26.6 149.3 17.8% County + Township 288.0 631.3 45.6% Total 658.8 2205.3 29.9% Source: Calculated from NDRC (2004) and CFY. Fiscal resources finance only approximately one-half of total PSU expenditures. The rest are raised from fees, donations, and business income. For example, although schools 49Here and in the following, the term "overstaffing" is used as measured against the government's own staffing plans. 38 for basic education are officially prohibited from charging tuition fees, until 2006 they were allowed to charge "miscellaneous" fees, textbook fees, notebook fees, and laboratory fees.50 Similarly, providers of agro-technical extension or cultural services raise substantial income from charging for their services. In some cases, these services might even be in direct conflict with their role as public service providers, such as when agro-technical extension agents sell fertilizers and pesticides to farmers. Likewise, health institutions receive very limited public funds and raise most of their revenue from selling drugs and charging for the services they provide. Today the relative shares of budgetary and non-budgetary funding vary substantially across PSUs by sector and by locality. Some data are presented in Table 4-3 for different types of PSUs in 2002. On average PSUs drew 48 percent of their financing from budgetary appropriations, but this ranged from 70.8 percent in the sports sector to 11.4 percent in health in 2002. Education was the largest sector, which absorbed 26 percent of the total budget appropriations for PSUs. These figures have changed since 2002, but the overall picture of great reliance on non-budget resources and the large variation across sectors persist. Table 4-3: The Revenue Status of PSUs in 2002 PSU revenue Share of financing from (RMB billion) budget National Totals 1,383.2 47.6% Education 357.8 59.9% S&T 75.0 56.9% By sector: Sports 6.4 70.8% Culture 21.1 49.1% Health 262.7 11.4% Central 240.2 48.6% Provincial 346.4 44.4% By tier: Prefectural 262.2 38.1% County 490.2 53.3% Township 44.2 60.1% Source: NDRC (2004). The large share of self-raised funds is not necessarily a reflection of inadequate budgetary support, however, since PSUs have a strong incentive to raise own funds to top-up staff remuneration. Data from NDRC (2004) show that total budgetary appropriations grew by more than 30 percent each year in both 2000 and 2001. The trend has likely continued and even accelerated, at least in support of basic education, as indicated by data presented earlier in Table 2-4. However, the increase in public funding was in most sectors wholly absorbed by increases in personnel expenditures. For all PSUs, total personnel expenditures grew faster than budget appropriations, a phenomenon that was especially marked at the county and township levels (World Bank 2005a). Examination of a national dataset for PSUs in World Bank (2005a) revealed that across provinces, staff remuneration was strongly correlated with the availability of self- raised revenues (see Figure 4-1). 50To impose some order, in recent years these fees have been rolled into a few categories -- miscellaneous, textbooks, notebooks, and boarding fees ­ and provincial governments have imposed some ceilings on their levels. 39 Figure 4-1: Plot of Average Pay against PSU Own Revenues by Province in 2002 60,000 50,000 40,000 pay 30,000 eragev A 20,000 10,000 0 0 000 ,000 0,000 000 20, 40,000 60,000 80 10 120,000 140, own revenues/staff Source: World Bank (2005a). There is justification in economic theory to charge users for the costs of providing most public services when the benefits are localized, the beneficiaries can be clearly identified, and the costs apportioned, and provided that charging for the service produces no significant adverse distributional consequences. However, levying user charges, although straightforward in theory, is very difficult to do well in practice (Bird 2001). Regulation of fee income requires a high degree of sophistication. In China, where the institutional environment is complex, and the economy is undergoing extraordinarily rapid growth and market transition, establishing a system of well-designed user fee regulation is not an easy task. A good example in the health sector is presented in Box 4- 3. To minimize distortions and prevent abuse, financial incentives have to be regulated by the government, sometimes through specialized agencies. Box 4-3: Price regulation in China's health sector Revenues of hospitals and other health care providers come from three main sources: out-of- pocket payments by patients, payments by insurers, and budgetary subsidies from government. The prices paid by both patients and insurers have long been regulated for most providers. However, despite many rounds of reforms, many wrong incentives still exist. Although the system of price regulation creates strong incentives for providers to deliver more drugs and sophisticated services, they face few incentives to deliver simple and cost-effective clinical and public health interventions. Not only are the prices of most basic services set below cost, providers at township and village levels are often expected to undertake population-based interventions, such as health promotion, school health, and water and sanitation improvements, for free. It is true that providers receive some budgetary subsidies from government. Ostensibly, these subsidies are paid at least in part to compensate providers for the difference between the regulated prices and the actual costs, and to support public health functions. In practice, however, they tend to be allocated using criteria such as number of staff and retirees, or the number of beds, with little or no regard to specific objectives or performance criteria. Moreover, government subsidies to providers account for a small and decreasing share of provider financing--ranging from just over 10 percent in the case of THCs to less than 5 percent for county hospitals. Hence, on the margin, government subsidies do little to encourage hospitals, THCs, and village clinics to make public health and basic services a priority. 40 Finally, current financing arrangements do not promote an appropriate "division of labor" across different levels of the health system. Indeed, with the current pricing structure, the provision of drugs and at least some profitable services are essential for survival. This explains the strong desire for THCs to invest in in-patient and high-tech capabilities. What is more, providers have disincentives to refer complex cases (which are more likely to be profitable) to higher levels. It is clear, then, that new ways must be found to finance health care providers. A reformed system should provide stronger incentives for cost-consciousness and appropriateness of care. It should also ensure that providers make basic, but highly cost-effective care and interventions a priority, and it should promote a clearer and more appropriate division of responsibilities among different levels of providers. In China the quality of user fee regulation and oversight for rural PSUs is low. In OECD countries, where local governments and service providers often rely on user charges for a significant portion of revenues, the charges are collected under strict laws and regulations that govern the level, scope, and use of the revenues, and the charges are subjected to periodic review and adjustment. In contrast, China has liberalized financial control over PSUs without introducing the requisite supervision and regulatory framework. With PSUs being given substantial freedom to charge for services and offered the incentive that they can "top-up" salaries through revenue-mobilization, the "pushing to the market" reforms of PSUs inadvertently replaced "service orientation" with "profit motives." The adverse effects are most clearly seen in the health sector, where health institutions focusing on making money often under-provide basic services ­ where government price controls keep the prices below cost ­ but sell too many drugs and high-tech diagnostic tests. Many PSUs have even exceeded their mandates and moved into areas outside their core functions. Some have also abused their monopoly positions to levy extractive fees that amount to quasi-taxes. While continuing to draw budgetary support, many PSUs have become highly commercial-oriented, turning fee- raising into an instrument for boosting staff remuneration. 4.3. Personnel Management of PSUs Labor is the most important factor in most PSUs, and personnel expenditures account for a substantial share of their operating costs, yet most personnel decisions are made outside of PSUs, by various government agencies, mostly at the level of the county government. The situation is similar to that for financial management, with responsibilities scattered across different departments of the local government -- the Post and Establishment Office, the Bureau of Personnel, the Bureau of Finance, and the relevant line agencies. Staffing levels are authorized by the local Office for Post and Establishment (at the same level of government). The overall number of posts allocated per locality and function is based on planning criteria that include many non-financial considerations: population size, size of territory, administrative status (level) of the unit, etc. Once hired, PSU employees earn life-long tenure after a short period of probation (usually one year). The Bureau of Personnel sets the framework for recruitment and retention for PSU staff and their remuneration. Roughly speaking, there are two categories of PSU employees: formal employees and "self-financed" staff. Formal employees have posts 41 that are authorized, and their salaries are supposed to be fully budgeted (although not necessarily fully paid) by the government. "Self-financed" (zishou zizhi) staff are, as the term suggests, paid by the non- budgetary resources of the PSUs themselves. Their posts are not authorized by the local Post and Establishment Office, and their salaries are supposed to be fully financed from the self-generated revenues of the PSUs. The formal distinction between the two categories is that the services provided by self-financed staff might be considered to be commercial in nature, rather than of a "public benefit" nature. In practice, this category is often used to take on extra staff, sometimes at the request of local governments, because the PSU has sufficient self-raised funds to pay for them. Because they do not require an authorized post, this category constitutes a "soft window" for expanding public employment. A recent study by the Ministry of Finance found the number of "persons supported by the budget" () ­ most of who are in local PSUs ­ continued to increase at an annual rate of nearly 4 percent per year since 1998.51 Employment terms for non-formal PSU staff vary by region and sector. However, until recently, the vast majority had de facto life-long contracts and received the same pay as the formal employees. Overstaffing is common to almost all regions and generally more severe in the less- developed regions. The number of staff employed above the quota of established posts is generally referred to as "overstaffing." Whether these are indeed "excess" staff is a matter of debate and depends on the definition one wants to apply. However, given the severe financial limitations and the seemingly excessive personnel costs vis-à-vis the operational funds of PSUs, overstaffing might indeed imply an "excess" over what is affordable. Overstaffing can be quite dramatic. For example, the recent reform in Hubei province (see below for details) suggests that rural PSUs (excluding those for regular education and health) in this central agricultural province were probably overstaffed by 60 percent before the reform.52 Remuneration for PSU staff ­ at least for the formal employees with authorized posts ­ is the same as that for civil servants in most respects, including medical and pension benefits, which are far more generous than those for enterprise employees. PSU wages mimic civil service wages and go up with the across-the-board wage increases that are periodically mandated by the central government and that affect all civil servants nationwide. It is only their greater freedom to earn informal salaries and bonuses that differentiates PSU staff from their civil service counterparts. Recruitment of PSU staff has largely followed the same rules as those for civil servants. Studies of PSUs have often found high proportions of unqualified staff.53 Low staff quality may also be because inland localities, such as Dayu County of Jiangxi Province, face many difficulties in recruiting qualified new university graduates to upgrade their public workforce. Dayu County has not been able even to fill all of its 51See Li and Xu (2006). 52According to data provided by the Hubei Provincial Development and Reform Commission, the abolition of township PSUs involved 91,211 employees. Under the new regime, only 36,701 (40.2 percent) succeeded in the competition and received service contracts. 53See World Bank (2005a) and unpublished background studies for this report, e.g., on agricultural and health services. 42 established posts and reported that during 2003-2005, 20 percent of its new recruits left before the end of their first term.54 Since 2002, the central government has been implementing a reform aimed to replace the implicit life-long employment with a term employment contract, known as "ping yong zhi." This is in response to the issue that the de facto life-long employment of traditional PSUs constitutes a major obstacle to almost all reform initiatives and performance improvements, as it makes it difficult to adjust the skill mix of PSUs to adapt to demand. Nationwide, 51 percent of PSUs and 58 percent of the PSU labor force have so far adopted the new regime.55 In particular, employees recruited since 2002 have been put under the new system in probably all 51 percent of the PSUs.56 This creates a precondition for a more flexible employment policy for PSUs that could be implemented in the future. The central government has also launched a new round of PSU compensation reform.57 This round of reform will place PSUs in a compensation system that is substantially differentiated, for the first time, from that for civil servants. Employment contracts and clarification of job descriptions are preconditions for the working of the new compensation scheme, which divides the total salary of a PSU employee into four components: three salary components based on (i) the job held by the employee, (ii) seniority and grade, and (iii) performance, followed by a component of subsidies and allowances. There are several challenges to implement this reform in rural PSUs. The first is the determination of the level of the total salary, as some PSUs have a high level of self-generated revenue, which typically translates into high and non-transparent actual pay. The second is the level of autonomy that can be granted to county government and the boards of devolved PSUs with regard to the length of the employment contracts for different categories of employees, the determination of performance-based salaries and subsidies and allowances. It seems recommendable for the central government to set up separate guidelines for implementation of this reform in rural PSUs, as these two challenges may be greater and more difficult to cope with in urban PSUs due to the much larger scale of revenue-generation activities and much higher actual pay granted to turban PSU employees than to rural PSU employees. A critical issue for any significant PSU reform is a proper arrangement to secure pension benefits for PSU employees who may have to be laid off and take up non- PSU jobs. Pension benefits for PSU staff are very generous compared to those for state- owned enterprise employees (see World Bank 2005a). Without addressing this issue, the de facto life-long employment and adjusting the structure of skills and the accumulation of overstaffing will be hard to solve. 54Officials cited a recent experience where they found themselves competing for new graduates from Jiangxi Normal University with salary offers of RMB15,000 per year, against offers of as much as RMB50,000 from coastal cities such as Wenzhou. Programs have been created to address these issues. 55Data are from the Ministry of Personnel. 56For example, out of the 1,500 employees in public schools in Yuan An County, Hubei Province, 169 were recruited under the new system, and this part of the labor force is growing by 40 persons every year. All "old" employees have signed employment contracts with the government except for 15, who failed in competition for a post and were given a period of two years to find a new job. During the two years, the government will offer them 70 percent of their basic salary. 57The last three rounds of a similar reform were in 1956, 1985, and 1993. 43 4.4. Accountability of Service Providers to Local Governments International experience suggests that, in order to deliver services efficiently and responsively, spending units need to be given a clear mandate, assigned appropriate resources, and held accountable for delivering their mandate. When spending units are given resources but not held accountable for results, funds are often not used efficiently toward achieving their intended outcomes. But experience also suggests that spending units can only be responsive and efficient if they have a degree of flexibility. Balancing authority with accountability is a key challenge to manage the relationship with service providers. Improving the accountability relationship between local governments and PSUs in China is critical to improving performance in public service delivery. The current combination of a rigid personnel system, overstaffing and under-funding, and lack of competition does not create a supportive environment for optimal results and sometimes does not go much beyond the struggle to "feed the staff" (yang ren). While local governments often do make performance compacts with PSUs, several factors impede effective enforcement. First, responsibility is not matched with autonomy. The decision-making autonomy of PSUs is very limited. For example, a school principal or the head of a township Education Bureau is held responsible for compliance with the "education objective- accountability agreement."58 However, her ability to influence school performance is constrained by current financial and human resources. As discussed in the section above, most allocative decisions, including hiring and firing and promotion of teachers, lie firmly with the county government.59 This situation violates a fundamental principle of good management: accountability and authority need to go hand in hand. Second, lax regulation of self-raised funds conflicts with the objective of providing quality public service. As discussed above, almost all PSUs in China have both public funding provided by the government and "self-raised" funds from fees, charges, and other incomes. This creates an asymmetry between local governments and PSUs: while local governments are almost wholly dependent on PSUs to deliver services for which they (local governments) are responsible, PSUs are not wholly dependent on local governments (for funding). This asymmetry creates incentives for service providers that are not well aligned with the service provision objectives of the government. In practice, granting service providers the right to generate self-raised funds is often a result of the ­ at least implicit ­ recognition that service providers are not receiving the necessary public funding for mandated services. However, without adequate regulation and oversight, these funds are primarily used for the remuneration of PSU employees rather than being allocated to the best use for effective service delivery. Third, the monitoring and evaluation system of local governments for the provision of public services has significant weaknesses. The shortcomings are similar to those observed at the central level. Too much data are focusing on inputs and outputs rather 58The system generally extends down to all teachers. 59As in almost all aspects of PSU management, significant variations exist across sectors and regions. For more details on education management, see Liu (2007). 44 than performance indicators; too much data are collected through the implementing line agencies, without the sharing of information with the public or even across line agencies, thus reducing the reliability of the data; and there is no tradition of county governments doing or, better yet, commissioning rigorous evaluations of their interventions. Fourth, competition between service providers is hardly existent. The creation of a competitive environment for service providers faces difficulties in many countries and, in particular rural areas. However, policy-makers in China could do much more to create a healthy competition on equal grounds between various providers ­ whether public or private. The issues mentioned above, i.e. clearly defined responsibilities, stricter rules governing service providers and their financing, and better M&E, need to be tackled as the more immediate efforts to create such environment. Fifth, enforcement relies entirely on personal responsibility. As noted earlier for local governments (Chapter 3), this mechanism is largely ineffective since the service providers are highly restrained by the fragmented management system that leaves key decision-making powers at the higher levels. What is required is a more systematic evaluation of the performance of service providers against their resources and autonomy. A good institutional set-up for public service provision depends on many factors and there is no universally applicable best model, nor a common formula to determine the best model. An improved accountability relationship between local governments and service providers can take many different forms, depending on the sectors and local circumstances. Determining the appropriate organizational form for providers of each service in a particular region will require a detailed review of the particular circumstances, a task best left in the hands of local governments. In fact, many successes can already be seen in the local pilot reforms, though many kinks remain to be worked out by the local governments. Nevertheless, the above analysis shows that the central government has an important role to play. Most importantly, it must set the regulatory framework that governs PSUs and other service providers, including their financing and personnel management, thereby setting the relationship between service providers and local governments. In addition, it can also play a crucial role in increasing the capacity of local governments to design and manage performance compacts with service providers. Improving service delivery will require a far-reaching reform process that should include a reconsideration of the role of the state and divestment from commercial activities, along with reforming the financing for public services. In the short run, allowing competition in the provision of public services -- either among government service providers or between government and private firms -- may bring some quick improvements in stimulating accountability. 45 5. CLIENTS AND CITIZENS Individuals are the recipients of rural public services and, therefore, at the center of the discussion on how to improve services. International experience suggests that enhancing downward accountability of service providers to citizens can in fact promote and complement their upward accountability to local governments. Chinese citizens' "voice" can be mobilized as an aid to nudge local governments to carry out central government initiatives for improving public services, even when they run counter to other local government priorities. A service-oriented government aims at meeting the demand of local citizens and communities for public goods and services. This demand, however, is not easily known since, unlike that for private goods, it is not straightforwardly revealed in markets, but only indirectly through political processes. Consequently, the "voice" and involvement of citizens and communities in the political process is crucial. In the terminology of the accountability framework introduced in Chapter 1, citizens and clients "delegate" to policy-makers and service providers the provision of quality services; they provide "financing" mainly, but not exclusively, through taxes; they obtain "information" as users of public services and through other channels; and they are critical for ensuring compliance. This chapter looks at the role of citizens, i.e. clients of public services have in this accountability framework vis-à-vis (local) governments and service providers. It discusses the processes in place for rural citizens in China to demand public services and participate in their design, delivery, and monitoring and evaluation of public services. It also discusses their role in the financing of public services. Finally, it looks at the organizations that enable the participatory processes. 5.1. Information for Public Services Provision The provision of quality public services crucially depends on the information flow between all key stakeholders. The previous chapters have discussed the importance of the monitoring and evaluation of the performance of local governments and service providers in terms of public service delivery. Several aspects of such an "information system" are of particular importance for this chapter. First, the performances of local governments and service providers have to be responsive to the needs, demands, and expectations of citizens and the services have to be assessed against their results for the clients. Consequently, the government and service providers have to obtain information from citizens on their priorities, preferences, and assessments of services delivered. Second, information flow is not a one-way street and the sharing of data collected and analyzed by the government and services providers with the citizens is central to efficient and effective public service provision. Third, citizens can play a very constructive role through their active participation in the monitoring and evaluation of services. These three aspects of information collection and flow will be discussed in the following. Information from Citizens to Government and Service Providers The government and service providers need to monitor the needs of and access for as well as the impacts on the users of public services. There are many mechanisms 46 and processes through which the various stakeholders can obtain this information. Obviously, government officials are influenced in their decision-making by talking to citizens directly, by looking at research studies, etc. More formally, China has traditional feed-back mechanisms through its letters and complaints offices. In addition, formal mechanisms exist for specific sectors and in specific locations. Nevertheless, most of the information of government and PSUs on users' needs and satisfaction is limited, and formal feed-back mechanisms on citizen satisfaction and impact of public services are weak in most locations. Surveys of public service users are an important instrument. These can be in the form of regular household surveys or score card systems. The household surveys undertaken by the Statistical Bureaus can be a basis for such instruments. However, more work could be done to improve the reliability of data and to standardize it for benchmarking, including for international comparisons.60 Although we are not aware of any survey that can claim to be nationally-representative, there are a number of studies that have tried to make such an assessment based on smaller sample surveys, often complemented by information through informal and formal discussions with the various stakeholders.61 Information about public service provision from users needs to be collected and evaluated systematically. Individual surveys can provide some independent view on the need and quality of public services and can contribute to raising new issues and providing advice in specific areas. However, they cannot compensate for a comprehensive and institutionalized system that regularly provides government as well as PSUs and other service providers with relevant information determining their performance. Many of the in-depth assessments have to be conducted for each village, township, and county separately and periodically. Information from Government and Service Providers to Citizens/Clients Providing quality information on public services to citizens is a cost-effective way to improve these services. Information on service needs or on services provided and their quality is crucial for central and local governments as well as for service providers to ensure that their policies, programs, and other interventions have the intended benefits. However, the information potential is not realized if it is limited to only within the system of government and service providers. The publication of information can be extremely effective by changing the behavior of citizens as well as that of local governments and service providers. Box 5-1 describes this for selected international examples of information campaigns. Box 5-1: International experience on information campaigns to improve services The power of information among citizens has recently been explored in several countries to improve rural services. Early evidence of the promise of this approach comes from the U.S. at the time when radio broadcasting technology had just been established. Between 1933 and 1935 60The World Bank is assisting in the assessment of China's education quality through internationally standardized surveys; it has also assisted the central government in monitoring and evaluating public service quality in urban areas. 61For an example see Zhao Yang et al. (2006) or DRC (2006). Given time and resource constraints, this study has not been able to conduct its own survey of rural public services. 47 federal assistance in the U.S. to low-income households was greater in those counties where more households had radios and were thus more likely to be informed about government policies and programs. The spread of radio improved information access, particularly for rural voters, who were previously disadvantaged relative to urban voters (since the latter already had access to alternative sources of information such as newspapers, while radio airwaves made it easier to deliver information to remote areas). It accounted for as much as 20 percent greater allocation of social assistance funds to a rural county as compared to an identical urban county. The politics of patronage or "clientelism" in the state of Ceara in Brazil was tackled head-on through massive information campaigns by the state government that took office in 1987. In only a few years this government tripled the coverage of measles and polio vaccinations to 90 percent of the child population and there was a fall in infant deaths from 102 to 65 per thousand births. The state government flooded radio airwaves with messages about how infant and child mortality could be drastically reduced through particular public programs by municipal governments, thus bringing political pressure to bear on the mayors to actually deliver basic health services. The state also created a new class of public health workers through a publicized recruitment effort that conveyed information to communities about the valuable role workers could play in improving public health through community-wide efforts. Such information campaigns have been credited with bringing a remarkable turnaround in the politics of the state--from being "clientelist" and patronage-based to becoming service-oriented. Another increasingly used tool is that of "citizen report cards," which compile information on citizen perceptions of the quality of services provided by various government agencies. These report cards are then publicized through local media and meetings with service providers. This tool was successfully used to improve municipal services in the city of Bangalore in south India. It served as a massive, collective complaint to the public providers that either "shamed" them into improving services or sent a strong signal to local politicians that citizens care about service delivery. In turn, the politicians put pressure on the providers to improve performance. However, such a report-card exercise for primary education services in rural India has not been successful. In contrast, a similar initiative in Uganda in primary health has substantially reduced infant mortality and improved service provision by local health workers. Sources: Banerjee et al. (2006); Bjorkman and Svensson (2006); Paul (2002); Stromberg (2004); Tendler (1997). Government mandates to disclose information in China has increased substantially. Many regulations today require public hearing meetings for policy formulation, particularly for policies closely related to public benefits. Broad-based information campaigns are becoming a more common practice in China and cover a wide range of areas, from administrative business at the village level to some policy formulation at upper levels of government; for instance, information bulletin boards in villages and townships disclose medicine prices in hospitals and clinics at the township and village levels. Many different instruments are used in the process of the information campaigns, such as public meetings, the media, information booklets, bulletin boards, etc. Information campaigns have led to positive results in many cases.62 The adoption of a new State Council regulation on government information disclosure shows that China is 62 For instance, a village committee disseminated project information to villagers through village meetings, information bulletin boards, and project information booklets for a health project, and this played an important role in mobilizing local people's participation and raising awareness of public health activities in a project financed by Plan International in Shaanxi (Zhang M. 2006). 48 determined to further improve information flow and, thereby, transparency.63 Proper capacity-building for the monitoring of local governments is critical to ensure enforcement of this policy. The rural-urban information divide remains large. The provision of information and communication services to rural areas itself is an important public service for the rural areas, but it is also important for improving the provision of other public services to rural areas. Significant investments have been made in the development of information and communication technology and the development of e-government, and the State Informatization Office has established overall guidelines for China's rural informatization.64 Implementation of these national guidelines will be crucial, but also a challenge, for the central government. In summary, strengthening the information flow to and from citizens is crucial for improving accountability in public service provision. China cannot rely only on "upward" accountability of service providers to local governments and of these to the governments at the provincial and central level. These accountability relationships have their limitations. Just as importantly, "upward" accountability is largely complementary and, in fact, depends on functioning "downward" accountability to citizens. Active Citizen Participation in Monitoring and Evaluation Involvement of citizens in the monitoring and evaluation of public service delivery can further improve efficiency in a cost-effective way. Such approaches, often referred to as "Participatory M&E", vary widely and must be designed according to specific circumstances. While conventional information collection and analysis is undertaken by governments and service providers, often by commissioning external experts, Participatory M&E involves the users of public services as the primary stakeholders and active participants (see Table 5-1). It thereby offers new ways of assessing and learning from change that are more inclusive, reflecting the perspectives and aspirations of those most directly affected. Table 5-1: Comparison between Conventional and Participatory M&E Conventional M&E Participatory M&E Local people, project staff, managers, Who plans and manages Senior managers or outside and other stakeholders, often helped by the process: experts a facilitator 63See http://news.sohu.com/20070117/n247677048.shtml. The regulation, adopted on January 17, 2007, requires governments at all levels to pay great attention to government information disclosure and to disclosure of information that is closely related to the people's benefits as the key content of government information disclosure. 64The guidelines call for (i) developing a rural telecommunications infrastructure (led by the Ministry of Information Industry); (ii) addressing the urban-rural divide by breaking the asymmetry of information, including that regarding government policies [...]; (iii) promoting county-level special industry/economy development [...]; (iv) developing county-level e-government in rural areas; (v) facilitating rural cultural development [...]; (vi) delivering social services using ICT applications in areas such as education, health, social protection, labor markets, family planning, etc.; and (vii) improving the capacity to absorb knowledge by the young generation in rural areas. 49 Conventional M&E Participatory M&E Role of "primary Design and adapt the methodology, stakeholders" (intended Provide information only collect and analyze the data, share beneficiaries): findings and link them to action Externally-defined, mainly Internally-defined indicators, including How success is measured: quantitative indicators more qualitative judgments Approach: Predetermined Adaptive Source: Institute of Development Studies (1998). Application of Participatory M&E is still limited in China. Obviously, the application of Participatory M&E varies widely by sector and specific local circumstances. Generally, however, it has been widely applied in international development interventions and a range of tools and methods exist for carrying out Participatory M&E. Participatory M&E is also applied in some domestic programs, such as the community- driven development (CDD) pilot program and the New Socialist Countryside campaign in Ganzhou in China, as shown in Box 5-2. But it is not yet applied in a systematic way. Nevertheless, the government has taken many steps in this direction and the central government has mandated to further strengthen this aspect of monitoring and evaluation. For example, Chapter V (1) of the 2004 State Council Decision on Reform of the Investment System states that "[a] social supervision mechanism for government- invested projects shall be established and the public and press shall be encouraged to supervise government-invested projects." Box 5-2: Participatory M&E in Ganzhou's New Socialist Countryside campaign Local people are well mobilized in the New Socialist Countryside campaign in Ganzhou. Among the different organizations at the community level, a monitoring group was established (see the diagram below) in Longshijian village. The main responsibilities of the monitoring group are to review financial income and expenditures, to ensure the publicizing of all income and expenditures in the village, to audit the use of community resources for priority needs identified by the local people, and to evaluate the performance of members in other community organizations, as shown in the diagram below. Two groups and two committees at community level Two groups Two committees Party group Monitoring group Village executive Consulting committee committee Working group for Working group for Working group for cultural activities Harmonious livelihood community development Source: Field visit to Dayu County, Ganzhou Municipality, Jiangxi Province. 50 5.2. Consultation and Active Participation in Public Services Provision The government increasingly is using consultation mechanisms in a wide range of areas. 65 Consultations are often a policy requirement for decision making and implementation of programs on potential controversial issues, such as land acquisitions and the demolition of housing, but it also is used for improving service delivery. For example, the Feasibility Study Guideline for Investment Projects in China, issued by the State Development Planning Commission in 2002, requires consultation with the intended beneficiaries for large- and medium-scale investment projects (SDPC 2002). Consultations are carried out in different ways for different purposes. Instruments include, for example, rural participatory appraisals, focused group discussions, open meetings, individual interviews, household surveys, questionnaires, and so on. Long- term consultation processes are institutionalized in organizations such as the School Development Committees.66 Similarly, various village community organizations67 were established as part of the efforts to build the New Socialist Countryside in Ganzhou Municipality, Jiangxi Province and played an important role in the consultation with and mobilization of local people. The further development and broad-based implementation of consultation mechanisms should be further enhanced. Local governments and service providers often do not use a consultation mechanism when reforming public service provision. This reluctance, despite substantial encouragement by the central government, may be caused by conflicting incentives. There is also a lack of local knowledge and experience about when and how to use different consultation mechanisms for different purposes. The most intensive method of interaction between government or service providers and citizens is through active participation. Active participation is a process by which stakeholders collaboratively define priority needs for public service, identify issues, develop options, negotiate solutions, and monitor and evaluate outcomes. In essence, it means the sharing of control over interventions traditionally decided solely by governments or service providers. The degree of active participation depends, of course, on the service sector and the specific local circumstances. Internationally, active participation is most frequently applied in development programs for small-scale infrastructure and the provision of public services at the community level. Successful programs in Indonesia demonstrate that such kinds of approaches can be applied in relatively large-scale programs in terms of their investment amounts and geographical 65 There is no strict delineation between information and consultation and a simple two-way flow of information might already be referred to as consultation. However, often consultation is referred to as an iterative process through which feedback is provided to decision-makers who incorporate it into public service planning and provision. 66 Such School Development Committees, which have been established in some village schools under basic education projects financed by the World Bank and other donors, normally include community members, teachers, and students. These organizations consult with local people about their needs for basic education and provide suggestions to school management to improve education quality. 67 The organizations include village executive councils, village consulting committees, community monitoring committees, and ­ supported by the Civil Affairs Bureaus ­ the model of "one association and five stations" (one volunteer association and five stations for social relief, health management, cultural management, dispute mitigation, and public welfare). 51 scope. Active participation is also applied in the provision of public services in other countries (see Annex 9). In China active community participation in rural public services is still in its infancy. It is, however, developing quite rapidly in certain areas and for certain aspects or stages of service delivery. For example, active community participation approaches are being tried in poverty alleviation programs.68 Moreover, some small-scale CDD pilots are being implemented; these pilots promote a particularly active form of participation by providing local communities with elements of direct control over resources, decision making, and management of funding and their own development activities. The process followed for these approaches is crucial. Annex 8 provides an outline of such a process for a CDD pilot in China. As is the case of other forms of participation, there are major obstacles to an accelerated replication of such approaches across rural China as local governments and service providers are not interested because of "loss of power," but at least as important is the lack of knowledge and experience. Nevertheless, increased participation is an essential part of changing the role of the government to one of providing services in response to local needs. 5.3. Financing Public Services The dominant involvement of rural citizens in the financing of public services is their direct financial contributions. Rural citizens in China pay directly part of the costs for most public services provided. Contributions range from education fees to payment for medical services, including public health, to labor or other contributions for the construction of small-scale infrastructures, such as village roads or the local water supply. These contributions range from being purely voluntary to being entirely mandatory, depending on how much participation and decision-making power is given to the citizens. This is shown in Figure 5-1. Voluntary donations are simply given for some activities of common interest. Other contributions might be induced by the government by providing only part of the costs for an investment or service and targeting communities that are able to mobilize citizens' contributions to match the funding gap. The range from induced to mandatory contributions is continuous. Induced contributions are usually associated with household-based investments, such as a biogas or a sanitation toilet program; user fees are collected from those using a particular service; and the contributions become "more mandatory" if the individual has no choice about using the services, such as in the case of compulsory education or the requirement to wear a school uniform, or if they are bound by social pressure or a collective decision to access a public good or investment, such as a contribution to match government funds for the construction of a village road. 68The new Poverty Alleviation and Development Outline also clearly points out that planning will use a participatory method, which takes poor villages as the base, defines the goals of the struggle, contents of construction, practical measure support units, and sources of funding. Currently, participatory village poverty alleviation planning has been carried out in 148,000 poor villages (LGOPA 2003). However, parallel reforms that shift financial and other control of village affairs to townships are in conflict and undermine these efforts. 52 Figure 5-1: Direct Financing by Citizens More community-driven mode Voluntary Induced User fee or Mandatory Enforced donation contribution service charge contribution contribution More government-driven mode Source: Authors. Certain financial contributions by rural citizens increase efficiency and should be encouraged. In particular, user fees for specific services can greatly increase efficiency and ensure that services can be delivered according to high standards. A prime example is the user fee for water based on volumetric measurement, which is directly related to consumption and saving of water. In fact, the building of the New Socialist Countryside explicitly calls for the raising of non-public funds. Such funds from beneficiaries and outside sources are believed to be a reason for the success of the New Countryside Program in South Korea. Depending on circumstances, disadvantages such as reduced accessibility for the poor might outweigh the advantages. Generally, the overall contributions by rural citizens are still perceived to be too high, in particular compared with those by urban citizens. The actual expenditure by rural citizens on public services is higher than that by urban residents, at least relative to their income. For example, rural residents are quite often asked to contribute to construction of small community infrastructures and schools, but urban residents are seldom asked to make such contributions. More urban residents are covered by a more comprehensive medical insurance system and by a social safety net than rural residents.69 Similarly, a recent survey of the reform of compulsory education in Jianli County, Hubei Province (Pi and Li 2006) indicates that most grassroots officials believe that although the farmers' burden has been considerably reduced since the rural-fee-reform, the issue of who should pay for basic education has not been solved. If unresolved, there is a risk of "rebouncing of farmer burden". This is also the case for the building of the New Socialist Countryside, which requires substantial contribution from rural citizens.70 Often, these contributions are seen by rural citizens as mandatory or at least not completely voluntarily. The government should address the weaknesses in the transparency and overall system of raising direct funding from citizens and communities and the standards of when, how, and how much to charge for public goods and services should be further reviewed and revised. The link between citizen contribution to and their say over the allocation needs to be enforced. Mandatory or enforced contributions are not consistent with participation. For investment projects, the yishiyiyi tries to enforce this and the design and implementation 69For example, China Society Periodical (No. 3 (2) 2006) reports that about 29.6 percent of urban residents, but 65 percent of rural residents are not able to go to hospital when they it is medically necessary. 70 100 percent of the costs for investment in cultural facilities was contributed by the villagers themselves in some New Socialist Countryside pilot villages in Dayu County of Ganzhou, Jiangxi Province; about 50 percent of school investment costs was contributed by villagers in some Plan International-supported village school development in Pucheng County, Shaanxi Province. 53 of this process should be strengthened. The stronger and more representative citizen's role in the funding allocation, the more likely the outcome of service delivery will reflect the needs of citizens for services. Such financial involvement in the decision-making process can have different forms. This includes community-level decision-making over public resources or participatory budget processes at a higher (government) level. Pilots of participatory local budgeting are also taking place in China. Over the past ten years, Wenling City, in the eastern province of Zhejiang, has been developing a system of "consultative democracy" that has allowed citizens to ask about and express their opinions on subjects related to their interests, particularly capital construction, road building, and education. Recently, this experiment has been extended by subjecting the budget review process -- or at least part of it -- to public discussion. Box 5-3 below describes the case of the participatory budgeting process in Zeguo Town of Wenling City. Box 5-3: Participatory budgeting in Zeguo Town Zeguo is a town of 120,000 people and 130,000 migrant laborers. In 2004, its economic output reached RMB1.4 billion, thus making it one of the richest townships in China. In early 2005, the government, after listening to proposals from the local people's congress and the CPPCC, developed an initial plan of thirty capital construction projects, including building roads and bridges, environmental protection, establishing parks, and reconstructing the older parts of the city. It was estimated that these projects would cost RMB137 million, but the town had only RMB40 million to spend. In order to decide on the priorities, the Zeguo government randomly selected 275 citizens for a participatory budget process. The group of citizens was given material that explained each of the thirty projects, and a group of experts introduced the projects. Based on this, the participants ranked the projects according to their order of priority on a preliminary questionnaire. This was followed by a day of democratic consultation; 359 participating citizens were assigned to sixteen groups, which discussed the projects and then presented their most pressing concerns to all the participants. A second round of small-group and large-group discussions followed, after which citizens again listed the projects in order of their priority. The results of this second questionnaire differed dramatically from those of the first, as well as from the predictions of government officials. The government then prioritized the top twelve projects for implementation in 2005. Source: Fewsmith (2006). 5.4. Institutions for Participation The government is leading the efforts of intensity and type of citizen participation in the management of public goods and service provision. On the one hand, participation, even the simplest collection of information and its evaluation, costs time and resources. On the other hand, there are significant benefits to participation and insufficient participation leads to situations whereby local people's needs for public services cannot be effectively addressed; the objectives of public service delivery without adequate citizens' "voice" is unlikely to be consistent with local demands. the success of participatory processes in the provision of public services crucially depends on the local governments' willingness and capacity to guide and effectively regulate these processes, through the setting up of a conducive legal framework, providing technical support for communities and service providers, etc. The lack of 54 knowledge about "how to" encourage community participation seems to be one of the most crucial obstacles to participation in China today. Different forms of ­ and therefore institutions for ­ participation are appropriate under different circumstances. Encouraging different forms of community participation clearly has a potential as well as limits. The collection and exchange of quality information is a precondition for any good management; the publishing of information is extremely cost-effective for improving service delivery; and the participation of citizens in data collection might further improve feedback and possibly reduce costs. More intensive consultation mechanisms are highly cost-effective for the design of new programs of service provision or other reforms. Finally, active community participation in planning, implementing, and operation of public services in rural areas can be very effective for some services, but unsuitable for others. One crucial factor is how much technical/expert information is required for the service delivery. Table 5-2 provides a highly stylized picture of what degree of participation might be suitable for different sectors. Table 5-2: Community Participation at Different Stages in Different Sectors Social Small Education Health Culture Security Infrastr. Preparation I C A I C I C A I C A I C A Operation I C I I C A I C A M&E I C I I C A I C A Note: I: Information; C: Consultation; A: Active participation Source: Authors. The effectiveness of participation depends crucially on the institutional set-up, which in China is dominated by the administrative village. Community organizations can play a crucial role in giving voice and ensuring involvement of individuals in the provision of public services.71 By far the greatest role of communities in public service provision is that of the about 650,000 administrative villages;72 and China has made substantial efforts to strengthen strengthening the structure of administrative villages.73 Nevertheless, this role of the administrative village as a community organization also has limits for several reasons. First, in many cases villagers do not see the village leadership 71 A community can be defined as a group of people who live in close proximity to one another or ­ in the context of this report ­ who access or participate together in the provision of public services. The size of a community varies depending on the type of services. It can be a natural village, several natural villages, an administrative village, or several administrative villages, but it can also be the community of all water users within hydrological boundaries or the community of parents and their children going to the same school. In this sense, individuals can be members of several "service communities." Similarly, the manner and depth of involvement of communities also differ widely by service sector and country or locality. 72 The traditional organizational structure of administrative villages consists of the village assembly and the two committees, i.e., the Party committee with the Party chief at its head and the elected village committee including the village head. 73 At the legal and regulatory levels, the Party Central Committee and State Council issued Guidelines on Improving the System of Openness of Village Affairs and Democratic Management in 2004 (General Office 2004), which is one of the latest steps to build on the 1987 Organic Law on Village Committees. In many areas village affairs are strengthened through the creation of specific committees and oversight bodies to ensure transparency and accountability. Increasingly, the village organizational structures also promote affiliated associations or committees for specific services such as cultural affairs, education, etc. 55 as representing them. Involvement by the Party and local governments or other traditional local power structures reduces the degree to which the village organs truly represent the villagers. In fact, the ambiguity of administrative village organs as self- governing bodies on the one hand, and as agents to implement government policies on the other, limits accountability to the stakeholders. Second, the ability of the administrative village has been weakened through the implementation of the Rural Tax and Fee Reform as it has severely limited the funding of administrative villages as well as their ability to organize farmers.74 Third, the administrative village is not the appropriate level of governance for many public services for practical reasons. The technically and economically most efficient area to provide public services is smaller or larger than that of an administrative village; it might be a natural village, a township, a county, and so on. Often such ideal areas cut across jurisdictional boundaries. For example, the provision of irrigation and drainage should be done according to hydrological boundaries. Overall, the significance of community organizations in China is still relatively small compared to that in most other countries (Plummer and Taylor 2004). Other types of organizations exist in the communities. However, they either have a traditional top-down structure, such as the so-called "mass organizations," or they are still relatively new and are in the process of being established, such as the water user associations. This lack of diverse community organizations is an obstacle to efficient participation by public service users. It is partly caused by the central government's legal and regulatory framework as well as by the lack of understanding, capacity, and incentives of local governments to promote such organizations. Strengthened grassroots-level organizations by rural citizens have the potential to increase the participation of beneficiaries in the management of services and thereby to improve performance. It is crucial that these organizations are fully driven by rural citizens rather than by governments or service providers. Simply creating community agencies does not ensure that people will be inclined to participate in the manner required to improve service delivery, for instance, if they do not rank particular services high on their list of priorities or if they are uncertain about the potential difference that can be made through their participation (see Box 5-4). Although examples abound of endogenously arising community groups that successfully manage or monitor rural services, it is not clear whether an externally imposed policy of community participation can deliver the same results because of the potential problems of elite capture and lack of citizen interest in participation. Box 5-4: Pitfalls of creating community organizations The creation of community organizations is no guarantee of greater participation. There are at least two problems that can occur. One risk is that these organizations can be "captured" by the local elite who use them for their own benefit and systematically exclude poor and disadvantaged people. Another risk is one of general behavior--do people have the incentives and preferences to participate in community groups to improve public services? Any policy of setting up institutions for community participation likely requires specific advocacy and information campaigns, targeted especially to poor and otherwise disadvantaged 74The introduction of a new "institution," the yishiyiyi, took substantial power and gave villagers the right to vote on individual projects to which they were to contribute in kind, labor, or cash. 56 citizens, if it is to be successful in generating local action for the public good. There is an example from Uganda that shows that such campaigns in primary health led to increased participation and striking improvements in actual health indicators. However, there is a counter- example from the education sector in India, where even elaborate campaigns executed by a capable NGO did not succeed in raising participation and in improving learning outcomes. The difference between these two cases appears to lie not in the design or efficacy of the implementation of the campaigns, but rather in the underlying social preferences for collective action. Sources: Banerjee et al. (2006); Bjorkman and Svensson (2006) A general policy lesson, therefore, is that where a demand for local participation exists, it can be harnessed, including through a formalization of group action, as a force for improving service delivery. There is evidence from within China that villages that have pre-existing social groups, such as village temple organizations or ancestral lineage associations, are more likely to have better quality public goods (Tsai 2004). In villages that do not have such pre-existing groups, the external creation of local agency needs to carefully select an area of service delivery that is high on the people's priority list and must invest in extensive advocacy and information campaigns. 57 6. SUMMARY AND RECOMMENDATIONS The report has analyzed the role of the four key stakeholders in the delivery of public services in rural areas, the accountability relationships among them, and how the central government guides these roles and accountability relationships. This chapter summarizes the key findings from the above analysis and concludes with recommendations for consideration by the central government to improve public service delivery in rural China. 6.1. Summary The system of providing rural public services is in a process of transition. Providing public services in rural areas used to be largely a local affair involving mainly township governments, villages, and their residents. Over time, the central government has gradually increased its presence, leading the efforts to strengthen service delivery by imposing national standards for public services, increasing financial subsidies to local levels, and instituting monitoring and evaluation processes. In recent years, these efforts have also included reforming the administrative framework, and many new programs have been rolled out to increase the scope and broaden the coverage of rural public services, and the momentum is building. These recent efforts have brought substantial improvements, with many benefits already tangible. For instance, rural infrastructural services are improving rapidly, school enrollments are up, wubaohu recipients are enjoying a much higher standard of living, etc. Remarkably, all of this is being achieved while the farmers' burden has been substantially reduced. The government's continuing focus on strengthening rural public services is well justified. The huge rural-urban gap in services is well-recognized by the leadership. China has one of the most successful economies in the world and its urban areas are making rapid progress in the provision of public goods and services. Although the significant achievements in rural public service delivery are laudable, rural services continue to lag far behind both those of the urban sector and behind the needs in the rural areas, and the top leaders continue to call for narrowing the gap.75 Nationally defined minimum standards for service provision in rural areas, such as free compulsory education, basic health care, a rural safety net, access to roads, safe drinking water, and so on, are within reach, but still need to be fully implemented. At the same time, success is important, as the "window of opportunity" to invest in human capital and infrastructure for sustaining future development is closing, given the country's rapid demographic transition (World Bank 2006a). An additional challenge is that the demands for public services in rural areas are changing rapidly. Many rural areas are undergoing dramatic demographic change as a large and increasing share of their working-age population is migrating to urban areas, and this dynamic situation has brought substantial changes to the types of services 75This was most recently reiterated at the National People's Congress in March 2007. See Wen Jiabao (2007) and Jin Renqing (2007). 58 needed. For example, in many villages there are fewer school children, but a greater demand for early childhood as well as elderly care. As the population becomes more mobile, there are also new demands for health care and other services to be portable, or transferable between in- and out-migrating areas. Moreover, within-rural inequalities are rising (Ravallion and Chen 2004; World Bank 2007d), requiring a better targeting of services and improved social protection systems. The strengthening of rural services requires that the government adjust to these changes and build institutions and mechanisms that are capable of more flexible, effective responses. The government faces many opportunities but also substantial challenges to further strengthen the system of public service delivery. The opportunities stem from the country's resounding economic success, which has boosted government revenues to the extent that they are growing even faster than GDP. This fiscal prosperity has brought about a climate that is amenable to redressing inequalities in rural-urban incomes and services, as detailed in earlier chapters. The challenges facing the central government lie in strengthening the multiple accountability relationships in the long chain from national policies to the citizenry ­ that between the central and local governments, government and service providers, and with citizens. As discussed in this report, each of these accountability relationships is in need of substantial adjustment and reform to fit the new conditions of China's increasingly market-oriented economy. The two greatest challenges for the central government are to ensure that local governments ­ down to counties (and townships) ­ have the capacities (including resources) for implementing the national agenda for public service delivery and for building mechanisms to hold them accountable for results. Currently, many local governments, especially in the western and central regions, do not have sufficient resources. At the same time, the incentives of most county governments are not fully aligned with the central government's priority for the provision of rural public services. Central ­ Local Accountability The central (and provincial) government is increasing its role in the provision of public services to rural areas. It is broadening the scope and raising the standards of the services that local governments should deliver. Increasingly, this is being accomplished by rolling out numerous programs that provide a national framework, but permitting, and sometimes even requiring, further design by local governments for implementation at the local levels. The rapid increase and growth of these new programs sometimes is straining local government capacities, at least temporarily. The central (and provincial) government is also increasingly providing financing to local governments for the provision of services, such as the increased funding and subsidization of compulsory education, health care, social security, etc. ­ now all local government responsibilities under the current Budget Law. As a result, rural governments in China have experienced virtually a full swing of the pendulum in their management of public services, moving from a locally determined pace of development and spending during the 1960s and 1970s, to the present where the pace is increasingly dictated from the top. In this process other changes have also moved local governments from having substantial autonomy in revenue mobilization, remuneration (especially of teachers), investments in schools and clinics, 59 etc., to a situation where their actions are far more circumscribed. On the revenue side, too, much of local government funding today comes in the form of transfers earmarked for salaries, capital investments, or other inputs for specific public services. The weak intergovernmental fiscal system remains a fundamental obstacle to the central ­ local accountability relationship, as it has been over the past two decades (Wong et al. 1995; Wong 1997; World Bank 2002, 2007a, 2007b). Despite significantly increased transfers and improved equalization, the intergovernmental fiscal system still does not ensure sufficient funding to counties and townships in the western and central provinces. This is true for overall amounts as well as for specific programs. For example, to fund Rural Compulsory Education adequately will require some RMB100- 200 billion annually, compared to the RMB20 billion provided in the current five-year plan from the central government. When responsibilities exceed resources, it is hard to hold local governments accountable for delivering the services or to judge their performance failures. Local government incentives for revenue mobilization conflict with central priorities for strengthening service provision. The numerous central mandates on spending, combined with limited own revenues and a complex transfer system dominated by earmarked transfers, leave little flexibility in local budgets. This reinforces the strong desire for own revenues (both on- and off-budget) at the local levels, whose growth depends on economic growth and urban development. The local preference for spending on infrastructure conflicts with the central government preference for spending on rural public services. Monitoring and evaluation of public service delivery is still weak. China has a relatively weak information base for local governments and local services. Although substantial data are collected by the Statistical Bureaus and individual line agencies, the information is not adequately standardized and often not shared with the public or even among agencies, reducing its usefulness and reliability. Too much information is focused on inputs and outputs, while the monitoring of well-specified and agreed-upon performance impact indicators for services is still in its infancy. This weak information base hinders the central government in making allocative decisions across sectors, as the lack of detailed, reliable data makes it difficult to undertake comprehensive assessments of public service delivery across sectors and localities to identify gaps. An evaluation of "culture" is still largely lacking, including at the central level. The accountability of county governments for the delivery of public services in the rural areas is weakened by other aspects of the intergovernmental fiscal system. Aside from the mismatch between resources and responsibilities discussed above, the lack of a strong formal assignment of responsibilities and the certain murkiness that results from this also cloud the accountability relationship. Without a formal system of responsibilities for subnational governments, local governments in China do not have clearly articulated roles and functions against which they can be held accountable. Enforcement mechanisms for the delivery of public services are also weak. First, given the administrative structure, any enforcement mechanism of the central government is "filtered" through the provincial and prefecture/city levels before it reaches the counties. This reduces its effectiveness as each lower level of government has its own 60 objectives that might compete with those of the central government. Second, little use is currently made of financial incentives, e.g., in the form of more transfers or more flexibility in the use of funding resources, to ensure compliance by local governments in rural service delivery. Instead, enforcement mechanisms rely largely on the personal responsibility system, using the performance reviews of government officials. The system of personal responsibility seems to have been extremely effective in the past in enforcing selected objectives, such as economic development and family planning, when local government leaders were judged against relatively few policy objectives. Going forward, China is putting significant effort into aligning this system with the new development agenda that emphasizes not only growth, but also balanced development and social and environmental objectives. Service-orientation is entering the performance evaluation of local government officials as well. These efforts clearly should be pursued and further refined. However, although the system of personal responsibility might still work well for a few "mandatory targets" that can bring exceptional promotions or "severe punishments," it will prove to be a less useful tool ­ on its own ­ for the complex objectives associated with public service provision that calls for nuanced enforcement. Just as the use of non-additive, multiple criteria proved unworkable in planned economies in the past, the new multiple objectives will likely yield similarly disappointing results as the dominant enforcement mechanism for service delivery. Personal responsibility cannot overcome systemic constraints, and assigning too much weight to personal responsibility risks undermining the whole system. Where set objectives cannot be met because of constraints beyond the personal control of the officials involved ­ such as teachers being held accountable for limiting student drop-outs -- and are widely recognized and accepted as not achievable, the system could be undermined by cover-ups and falsification of data. The usefulness of the cadre responsibility system is further diminished in the current environment, where many more national policies and regulations are in place to limit the degree of freedom within which local officials can operate. In sum, the current trend toward greater central regulation of local services and spending has the important benefit of imposing more uniformity in public services across the nation, but it comes at the expense of losing the benefits of information and initiative that local management can bring. This trade-off is all the more significant given the highly decentralized administrative structure in China, where the central government lacks a capacity for detailed, top-down management and control of public services that are implemented by local government levels. Service Providers Just as the central government depends on local (county) governments to implement policies on rural public services, local governments depend on service providers to deliver services. For most public services, the vast majority of providers are public service units, which overwhelmingly dominate rural compulsory education and in-patient care hospitals and township health centers. Since PSUs account for nearly half of all budgetary expenditures at the county and township levels, improving the interactions between local governments and PSUs is obviously critical to efforts to strengthen rural 61 public services. At present, the accountability relationship between local governments and PSUs is marred by many of the same weaknesses as the central-local relationship. These include lack of clarity in responsibility assignments, insufficient funding, and weak monitoring and evaluation. The lack of clarity in the assignment of responsibilities to PSUs stems principally from the lack of clear boundaries between government and PSUs. The use of "letters of responsibilities," "education objective-accountability agreements," or other forms of performance compacts is relatively widespread. However, there are other mandates, some of which might not be an explicit part of the performance compact. For example, in addition to providing public services, many PSUs perform core government or administrative functions. Even if this is not the formal task of the PSUs, the use of PSU staff to support the work of government cadres is common practice in many areas. PSU heads have very limited allocative power. The greatest constraints to a stronger focus on service provision and the most important infringement of management autonomy in PSUs likely is the personnel system, which continues to be closely linked to that of the civil service. Some reforms are underway, but in general PSU staff have public employee status and acquire life-long tenure after a brief probation period, and their basic pay moves in step with civil service wages. Unlike SOEs, PSUs have little autonomy in recruitment and retention decisions. As many local governments continue to use PSUs as employers of last resort, assigning staff without regard to qualifications, the problems of overstaffing and unqualified staff in PSUs have both exogenous and endogenous causes. Many other inefficiencies in PSUs stem from the incomplete property rights assigned to them. Inadequate public funding for services weakens PSU accountability to local governments. In the aggregate, budgetary resources finance just half of the expenditures of PSUs. Even core public services such as rural compulsory education and maternal- child health receive only partial funding from the budget, and PSUs rely on fees and other sources for a substantial portion of their expenditures. The pervasiveness and importance of non-budgetary funding over time has eroded the link between public funding and public service provision, further weakening the accountability of PSUs to local government. Reforms over the past two decades have introduced market-like incentives to PSUs, in most cases without exposing them to market competition. Nor are PSUs subjected to strengthened financial control or accountability for performance. Allowing PSUs to use a portion of their "self-raised funds" to add a "flexible" component to remuneration boosted incentives for PSUs to mobilize non-budgetary revenues. Under the generally lax government oversight over self-generated revenues, many PSUs adopted revenue- maximization, rather than service provision, as their principal objective (World Bank 2005a). In sectors where market failures are significant (such as health and education), revenue-maximization has not only distorted the mission of many PSUs, but also has imposed excessive burdens on citizens and created problems of access for the poor. Citizens and Communities Participation by citizens and communities in service delivery is increasing rapidly. The amount of information provided and transparency has risen, with a positive impact 62 on the accountability of local governments and service providers. Even so, much more could be done to increase the usefulness of the information provided and to further extend ICT to rural areas, and Participatory M&E and feedback mechanisms could be strengthened. The level of consultation between local government or service providers and citizens or communities is also increasing, but institutionalized processes are still relatively limited. An even more active participation with decision-making power at the community level is still largely limited to certain programs and selected localities. Obstacles hindering further progress are only to some extent of a regulatory nature, e.g., for the promotion of community and civil society organizations. An even greater challenge is the lack of general understanding among local governments and service providers regarding the concept of participation and its benefits. The role of citizens and communities in financing public goods has been dramatically reduced with the rural-fee-reform. Nevertheless, rural citizens still contribute very substantially to public goods and services, and to a far greater extent than urban citizens. Despite improvements in attempts to enforce illegal school fees or excessive prices for medical services and in regulations such as the yishiyiyi, the financial contributions by individuals and communities are often excessive. Significantly, these contributions are not always matched with an influence on allocative decisions. A direct influence on budget allocations is largely limited to the village planning exercises for poverty reduction, and it is not common practice for the bulk of community-level funding. 6.2. Recommendations The transition of rural public services is incomplete. Many of the recent reforms remain to be fully implemented and consolidated. Adjustments are still needed in the assignment of responsibilities and finance; the personnel and M&E systems are not well aligned to ensure effective enforcement; public service providers are often inefficient, self-interested, and lacking in a service-orientation; and the potential role of citizens and communities still remains largely untapped. We therefore propose the following key recommendations for consideration by the central (and provincial) government: Clarify Delegation (vis-à-vis local governments) Simplify goals and focus on only a few strategic and achievable targets. The central government should harness its growing fiscal capacity to focus on a few key sectors such as rural compulsory education and improving public health and health insurance coverage ­ areas that are the core elements of the Harmonious Society and New Socialist Countryside ­ rather than spreading resources thinly over many areas. Given that local governments are responsible for implementation, the goals defined at the central level should be in terms of performance outcomes. Most aspects of implementation for achieving these goals can be left to the discretion of the local governments. Clarify and formalize responsibilities for each level of government. This might include assuming more responsibility for financing and perhaps even provision of some rural public services, e.g., social security and the NCMS, which are beyond the capacities of county governments (especially given migration). Making explicit and formal what 63 each level of government is responsible for is a precondition for a strengthened accountability system between the various levels of government. Ensure Financing Match responsibilities with finance. To enable local governments to fulfill their assigned responsibilities, adequate funding has to be assured. This can be achieved through a combination of increasing local own revenues and improving transfers. Local own revenues could be increased by providing local governments with more taxes and/or greater autonomy on tax rates, including for the property and resource taxes, and permitting them to levy surcharges on income taxes. A greater reliance on user charges should also be considered under strengthened supervision. The efforts to bring off- budget resources into the budget should also continue. Assigning greater revenue autonomy will not solve the problems of all localities, especially many in the central and western regions. More funding and greater equalization will also be needed in the transfer system to close the still existing fiscal gaps. Improve the efficiency of funding. The need for a revamping of the fiscal transfers system is well-recognized, including a consolidation of earmarked funds to broader sectoral transfers and a limiting of transfers for only specific elements of service delivery, such as salaries for teachers, equipment for health centers, etc.76 The central government should also take a bolder approach in budgeting to allocate resources fully in the annual budgeting process and avoid mid-year, incremental budgeting. In particular, for the (few) programs that are identified as top national priorities (e.g., universal basic education) the central government should provide full funding. Reform the Governance Framework for Service Providers An improved governance framework for PSUs should include the following elements: (i) a clear separation between core government agencies and service providers; (ii) differentiated treatment for delegated and devolved PSUs; and (iii) introduction of effective performance compacts between local governments and PSUs. Continue to increase the flexibility for employment by PSUs. The recent introduction of flexible, term employment contracts is an important step forward, and in the longer term reforms should aim at eliminating the post establishment system for PSUs, since the system has been ineffective in preventing overstaffing. It is also crucial to tackle the issue of pensions for retired PSU staff. At present, these are paid for by the PSUs themselves from current revenues. In reform pilots such as that in Hubei Province, the solution is generally to shift PSU employees from the existing budget-funded pay-as-you- go system to the Basic Pension system for enterprise workers.77 However, such changes 76Local budget discipline, budget allocation, and efficiency are also limited by personnel policies in terms of staffing and wage levels. Personnel policies are the most disruptive at the lower levels of government, where personnel expenditures absorb the bulk (and sometimes the whole) of budgetary resources. See point on setting the framework for service providers. 77Two important issues are involved in Hubei's reform. First, the government budget needs to pay the enterprise pension fund the contribution (both social pooling and individual account) that should have been paid had the employee been an employee of an enterprise. Second, due to the higher benefit PSU employees are entitled to as compared to enterprise employees, compensation must be provided to PSU employees who are close to retirement, typically through a special subsidy that equals 100 percent of the 64 require huge upfront costs that are unaffordable in many localities and PSUs, so an overall reform of PSUs will require a systemic solution with central government subsidies. Strengthen financial management of user fees and fee revenues. Such efforts should start with a clear and well-enforced code of conduct for public employees that strictly limits the use of fees for staff remuneration. The conversion of fee income or explicit financial incentives for PSU performance for the use of PSU personnel distorts PSU incentives and leaves insufficient funds for non-personnel expenses, which are also an important part of quality services. Strengthen capacity of local governments and service providers. The central government could significantly help in this area through the development of management tools, improved knowledge management, and the financing of training and technical assistance for local officials, such as study visits to learn how local governments in other countries have managed contractual relationships with NGOs and for-profit enterprises for the delivery of public services. Capacity-building efforts should also include upgrading the workforce of PSUs. For example, the central government has already introduced a new program to "purchase teachers' posts," funded with more than RMB200 million in earmarked funds.78 Jiangxi Province has a similar innovative program whereby 105 new university graduates were recruited in Ganzhou and sent to the township level for two years, with the province providing 30 percent of their RMB800/month salary and the county providing the rest. After two years, these recruits will be given PSU posts and assignments. The program is broadly defined, with recruits in teaching, medicine, and other fields. Increase competition in service provision. China has already begun experimenting in using non-public providers, such as "buying services with money" discussed in Chapter 4. Annex 5 presents another example of a pilot reform in Hubei Province that has attracted nationwide attention. The central government could provide a framework to guide considerations in areas such as (i) the appropriate public-private cost-sharing; (ii) the involvement of farmers and farmers' associations; (iii) the elimination of entry barriers in markets; and (iv) the organizational form of "service centers." Review and assess pilot reforms to ensure compatibility with the core goals of public service reform. A decentralized approach will be needed given China's vast size and regional diversity, and because the nature of public services differs greatly across sectors. Sweeping reforms as those in Hubei do not suit all localities or all sectors. A blank "push to the market" is certainly unadvisable. Prime candidates for greater private involvement in provision, with at least some public funding for what is perceived to be public goods and services, are cultural and agro-technical extension services.79 Issues gap for those who are retiring in one year, and a diminishing percentage for those who are retiring in two to five years. 78In 2006 some 16,000 recruits were enrolled, with the central government paying an annual salary of RMB15,000 to send them as teachers to the western provinces. The receiving governments commit to take these teachers on their own payrolls after three years and to pay for benefits, including social security contributions and medical insurance, and living subsidies. 79 These types of services can demonstrate a relatively higher degree of measurability and contestability, if farmers as consumers are properly empowered in M&E and the market is fully liberalized and well 65 facing schools and health clinics, on the other hand, are related more to their financing and internal management than to whether they are public or private.80 Nevertheless, the government could still try to encourage non-government participation in rural education and health to make the two markets more competitive and contestable. This would require a leveling of the playing field for service providers of various ownerships81 and could be encouraged through public-private partnerships in areas where conditions are ripe. Annex 6 provides two successful examples of this kind in Latin America. Encourage Citizen and Community Participation to Enhance Accountability Strengthen the regulatory framework and support capacity-building at the local level. Governments in many countries have found that citizen- and community- involvement in public services helps to promote greater efficiency and "demand- responsiveness." Participation is no panacea, however, and not all approaches are suitable for all services. Given China's political system, community participation cannot replace a strong upward accountability of local governments and service providers to higher levels. Nevertheless, participation can help to strengthen both downward and upward accountability of service providers. The role of the central and provincial government is largely limited to two areas: setting the right legal and institutional framework, and promoting the implementation of more participatory processes through awareness-raising and capacity-building. Increase information and transparency. China has been successful in improving this important area, which is crucial in all aspects and sectors of government and service providers. However, it could further strengthen disclosure rules and their enforcement as well as improve the relevance and presentation of information. Similarly, the media for information could be improved, in particular through the further strengthening of ICT into rural areas. Closely linked, but going beyond traditional approaches, is Participatory M&E, which makes citizens and communities the driving force in information collection and analysis and should be directly linked to decisions about what and how services are provided. Strengthen consultation and active participation. In many areas, participation should go far beyond information and transparency. Consultative processes can improve budget allocations as well as service provision in sectors such as education. Some community- level services such as many cultural services and village-level infrastructure can be most efficiently and effectively designed and implemented by the communities themselves, e.g., through so-called CDD interventions that recognize villages, natural villages, and regulated. The rationale for government provision of these services is generally weak, despite sectoral and regional variations and possible special cases. 80 Given the low measurability of education and health services, government ownership has the potential to contribute to performance improvements by promoting public service ethics, strengthening input and process control, and improving M&E that involves multiple stakeholders. 81For example, students enrolled in private primary and secondary schools should be supported by the same set of financing policies applicable for universal compulsory education, and the treatment of private health care providers in the New Cooperative Medical Scheme should be the same as that of public providers. In areas where private primary and secondary schools have significant potential to develop, it may be worthwhile to pilot an education voucher program, not necessarily to create competition but at least to ensure a level playing field. 66 other communities as legitimate levels for the design and implementation of public goods and services, including those with public funding. Establish effective standards and models for community financing in public services that link financial (other than taxes) or other direct contributions to substantial decision- making powers given to the providers of the finances and users of the services, by (i) increasing citizen and community involvement in the allocation of public funds; (ii) encouraging community financing for specific purposes on a voluntary basis by promoting a more community-driven mode; and (iii) fully publicly funding services in cases where affordability is an issue. Build a Nationwide Monitoring & Evaluation System The implementation of the above recommendations crucially depends on an effective monitoring and evaluation system, based on which the central government (and other stakeholders) enforce standards, policies, and regulations for service provision. For the central and provincial governments, this entails the following: Building a strong M&E system that enables it to assess performance of local governments vis-à-vis outcome objectives, as well as to evaluate the effectiveness and efficiency of policies and other interventions. Although China is moving in this direction, it could accelerate the process by several measures. First, outcome performance indicators should be decided upon and the collection of information at the central level should be shifted from the current focus on input and output data to this information. Second, transparency should be brought into the monitoring system by sharing and harmonizing data across line agencies and making it available to the public. Third, policy interventions should be systematically evaluated by requiring data reporting and regular evaluations ­ preferably by third parties. What these recommendations require is strong leadership from the highest level and a "champion" ­ outside the line agencies and closely linked to allocative decisions and other "rewards and punishments" discussed in the following. Strengthen Enforcement Continuing the refocusing of the incentive system of local government officials on (rural) public service delivery. The personal responsibility system that runs through China's Party, government, and PSU system overall is very effective in achieving key targets set by the leadership. The system could be strengthened to fully reflect the weight given by the country's leadership to the provision of public services, in particular in rural areas. Specifically, the central and provincial governments should continue their ongoing efforts to broaden the evaluation of local government leaders and officials from the traditional economic growth focus to a focus that more strongly reflects performance in public service delivery. Limiting the use of direct financial incentives for individual officials and employees. China appears to be unusual in terms of relying on strong financial incentives to motivate public employees. The governments of most other countries limit financial awards mainly to nominal levels, and instead rely on promotions and career prospects as key incentives. 67 Increasing the use of fiscal transfers to reward local governments for performance in public service delivery. Surprisingly, given the preponderance of transfers made by the central government, and in contrast to the extensive use of financial incentives to reward PSU staff, the central government makes little use of such incentives for local governments. Most fiscal transfers are provided on the basis of criteria such as taxes collected, number of employees, share of minority population, degree of poverty, and fiscal deficits, and are not linked to the performance of local governments in public service delivery. The amount of transfers that are earmarked for public services is also rarely dependent on local government performance, and tends to be allocated based on "needs" or other criteria. Given the huge size of the central transfers, this is an area of great potential to be considered for strengthening rural services. Strengthen Capacity Strengthening the multiple accountability relationships cannot be achieved without substantial investment in capacity-building. Efforts need to be led by the central and provincial governments. Given the tremendous heterogeneity within China, "horizontal" knowledge exchange and management between local governments and PSUs would be extremely effective ways of strengthening capacity. Awareness-raising, training, and technical assistance are needed in many areas, but budget and planning processes that reflect local needs in a changing demographic environment, participatory processes, and monitoring and evaluation (through capacity-building and adequate incentives) are likely to be the greatest bottlenecks. 68 7. OUTLINE OF A REFORM PILOT As discussed above, improving public service delivery in rural areas requires comprehensive solutions. At the same time, the Government might want to test key elements of a comprehensive reform strategy on a pilot basis. This Chapter outlines such an innovative and ambitious reform proposal that addresses some of the core issues for rural public service delivery identified above, that explicitly test various solutions against each other, and that can be carried out and evaluated on a small scale before being scaled up. Obviously, the pilot proposed below is only a first rough draft and needs to be elaborated in discussions with key stakeholders. 7.1. Overview This report has shown that China's system is reforming rapidly to improve public services in rural areas. Major efforts are being made in practically all relevant aspects. Some responsibilities are being redefined; the allocation of financing is changing; the monitoring and evaluation system is being reformed and strengthened; PSUs are being reformed and other forms of service providers are emerging; participation by citizens and communities is strengthening; significant reforms are being undertaken in education, health, and many other sectors; and so on. This picture shows how difficult it is to propose yet another reform initiative. Some reforms are already well-accepted and rolled-out nationwide. Others are still at an experimental stage, often pursued through local pilots. To some extent, the multitude of changes leads to some degree of "reform fatigue" at the local level as not all changes are consistent with one another and usually too little training and technical assistance are provided to the local governments and service providers for the implementation of these reforms. Most crucially, the benefits of many pilots could be enhanced through strengthening evaluations. A new reform pilot needs to take these realities into account. We therefore propose a pilot reform that (i) tackles a few of the most crucial bottlenecks for public service provision in rural areas; (ii) is sufficiently comprehensive in its design to effectively build on and be consistent with other reform programs; (iii) is sufficiently ambitious to have a substantial impact on service delivery, but can be tested in a limited area only; and (iv) focuses on areas in which the NDRC has an important or leading role to play. At the core of such a reform pilot we see the strengthening of two interlinked factors providing incentives to county governments: fiscal transfers for improved service delivery and the monitoring and evaluation system. This choice addresses the criteria outlined in the previous paragraph. First, this choice reflects the above analysis and tackles some of the key bottlenecks, in particular: (a) China's current accountability mechanisms rely heavily on a strong personal responsibility system. While the system is working well in many respects, as a sole measure to achieve accountability it is limited and it has drawbacks that are reflected in the channeling of financial incentives predominantly to 69 personnel expenses rather than to other expenses critical for service delivery. Complementing this system with a system more strongly focused on providing fiscal incentives for institutional accountability between various levels of government and institutions would strengthen the overall system. (b) Weaknesses in the monitoring and, especially, the evaluation of policies and programs clearly present a bottleneck to the improvement of public services in rural areas; in fact, performance indicators focusing on the impact of public service provision reflect the objectives of the building of the New Socialist Countryside, but are not collected and analyzed in a systematic way. Second, the pilot is comprehensive and can be consistent with other reform programs. Adequate monitoring and evaluation would take into account all relevant aspects of public service delivery and would assess not just the narrow pilot initiative, but other policy and investment interventions in a targeted county. Third, the pilot would have a fundamental impact on the formulation of policies and the manner in which local governments operate. The strengthening of the "institutional" accountability of local governments combined with improvements in the monitoring and evaluation system is clearly ambitious; at the same time, it can focus on one or only a few local jurisdictions and sectors. In fact, the proposed pilot is fundamentally different from most traditional pilots. While a traditional pilot would promote reforms, new initiatives, and would build capacity to improve outcomes, the proposed pilot primarily aims at improving the incentive system of local governments to provide quality public services in rural areas, and thereby to stimulate them to undertake reforms, new initiatives, and to build capacity for the improvement of outcomes. Moreover, while traditional pilots in China focus on testing a particular reform model before it is scaled up, the proposed pilot would include different solutions to aspects that are best solved `empirically'. The proposed pilot would test various models against each other through a different design in different counties and careful monitoring and evaluation. Fourth, the NDRC has a mandate to strengthen the M&E system and significant power over the allocation of fiscal resources to local governments. M&E efforts should ­ as international experience suggests ­ be led by a strong and comprehensive agency in order to ensure independence from the implementing line agencies and in order to effectively impact resource allocations. Moreover, the NDRC's ability to effectively lead and coordinate broad policy formulation crucially depends on its ability to access and evaluate reliable data on public services. It is therefore well positioned to further strengthen the already ongoing efforts to improve the M&E system. As discussed above, the NDRC's role in the allocation of fiscal transfers is mainly restricted to capital investments, but it has substantial allocative power through the planning processes it leads. Nevertheless, an ambitious pilot would require strong engagement by the MOF as well. The design of such a reform pilot will have to be undertaken in close cooperation with the participating localities and be assisted by some technical advice. It is also obvious that such a pilot as proposed here would entail some risks. The greatest risk of failure would be from insufficient leadership, resulting in insufficient coordination across 70 agencies at the central and provincial level as well as insufficient attention at the level of the participating counties. 7.2. Objective and Institutional Set-up The objective of the reform pilot is to contribute to the building of a New Socialist Countryside by improving the delivery of public goods and services in rural areas and, more specifically, to achieve minimum standards of service delivery in poor rural areas. The success of the pilot project would be measured in terms of the results or the impact on public goods and services (e.g., education enrollments, drop-out rates, quality of learning, etc.) and how cost-efficiently these results are achieved. The strong need for political commitment and leadership for the success of such a reform pilot would require very high-level leading groups at the national, provincial, and county levels. Under the leading groups, Program Management Offices within the NDRC and the provincial and county DRC would be formed. These offices would also include representatives from the MOF and local Finance Bureaus which would have frequent interactions with the relevant line agencies and would be responsible for the day-to-day implementation of the reforms in the pilot counties. In line with China's positive experiences of pilots, the pilot would be tested on a limited scale. Since the proposed reform pilot is relatively comprehensive and therefore requires substantial central-level monitoring capacity, it is recommended that the pilot be focused on a limited number of counties within only one or a very few provinces (in central or western China). A political commitment to such a reform is the key selection criteria for participating localities. The key implementing level of the pilot would be the county since it is the county that determines how national policies, funds, and other incentives are translated into actual service delivery to rural citizens. The full political buy-in of the county leadership is therefore crucial for the success of the pilot. A limited number of counties should participate on a voluntary basis after detailed vetting of the pilot design. The entire jurisdiction of the county, i.e., all its townships and villages, would be included in the pilot. The pilot could consider selecting counties on a competitive basis. 7.3. Design Process and Components A reform pilot is a process that needs to be well planned and implemented over a number of steps and years. The sequencing of the design and implementation steps and the availability of sufficient time for all stakeholders to adjust to the new incentive system are crucial for success. Once a general political willingness to undertake the pilot is assured at all relevant government levels, the specific objective of the pilot and the performance indicators need to be agreed upon. Setting the right performance indicators is no easy task, especially in an environment as dynamic as that in China where the quality of service delivery crucially depends on external factors, such as migration. It requires substantial analysis and discussion. Nevertheless, the establishing of a baseline for what should be 71 achieved in each of the participating counties is of crucial importance. The review of the baseline situation would also include an assessment of the participating villages and the townships' human and natural resources, and any trends thereof. Another initial step is the starting of improvements in the local-level planning and budgeting processes. This requires the establishment of an initial comprehensive account of the counties' (including the townships') fiscal and administrative resources and capacities, as well as of government policies and programs currently being undertaken. Such a review would likely result in reallocations of resources in line with policy objectives, as well as increased transparency ­ within the administration, the local people's congresses, higher-level governments, and, possibly, the public ­ and thereby accountability. An important component of the pilot would be the strengthening of the monitoring and evaluation system throughout its implementation. As stressed throughout this report, a better M&E system does not mean more data; rather it would improve (i) the type of data and the way information on rural public services is collected and reported, including to the NDRC; (ii) the way this information is used in rigorous evaluations; and (iii) the way the evaluation results trigger "enforcement," i.e., the impact of the policies and programs being evaluated. The component would strengthen and streamline the existing data-collecting mechanisms82 and commission the evaluation of various policies and programs in the participating counties, obviously including the pilot itself. The reform would provide the participating counties with financing and financial incentives for performance in terms of service delivery. This could include a block grant that might replace some of the existing earmarked transfers. Part of the funds would be provided with the stipulation that their performance in improving the indicators on which they are targeted would be monitored ­ non-compliance would require a further strengthening of capacity-building and, possibly, a shift in responsibilities. Various models exist83 and need to be carefully evaluated during the pilot design process. As several possible options might emerge, the pilot should adopt these different approaches in different counties and through implementation and careful monitoring and evaluation test them against each other. Another crucial element in the pilot is the provision of training and technical assistance, partly with guidance and financing from higher levels. Technical assistance might be targeted toward improving the budget process, giving more responsibilities to communities, and increasing participation; reforms of the PSUs; or any other topic,84 largely based on the demands of the participating counties. An important area is to 82 Complementing institutionalized systems with an independent household-based data- collection activity that is implemented by a higher-tier national agency or an independent institution might be valuable under certain circumstances. 83See, for example, the combination of expanded block grants and separate performance grants under the Bangladesh ­ Local Governance Support Project (World Bank 2006b), the Local Government Development Grant that stipulates minimum conditions to its access under the Sierra Leone ­ Institutional Reform and Capacity Building Project (World Bank 2004d), or the case of South Africa presented in Box 7-1 below. 84For elements of technical assistance provided to local governments by central programs see, for example, Kenya ­ Institutional Reform and Capacity Building Technical Assistance Project (World Bank 2005c). 72 increase the capacity for monitoring and evaluation at the local government level itself. Training would target county leaders, officials, service providers, and communities. It could be delivered through vertical channels, i.e., organized by higher-level governments and provided to the counties, and also through "horizontal" learning in which county governments or service providers or communities are linked. As holds for the other component, different approaches can also be implemented for training and technical assistance in different participating counties and tested against each other. The three key elements or components would be combined in a compact or "contract" between the county and the higher-level government, with sufficient specificity and viability. The stated outcomes of the plan would serve as the basis for the monitoring of agreed actions, outputs, and outcomes. The concrete design of the broadly outlined components depends on the scope of the pilot. This should be broad, but might be reduced to a specific sector, such as several or selected social services or rural infrastructural services only. The financial incentive for a county to participate in the pilot would be largely based on increasing the county's flexibility in terms of the specific (largely earmarked) transfers the central government provides for the targeted sector/area. For an example of a pilot that focuses on infrastructure only, the participating counties might receive an annual allocation of NDRC funds - of a total amount that might be based on past allocations - at the beginning of their budget years. The allocative decision regarding which type of infrastructure should use these funds, and whether to use the funds for new construction or renovation, would be left to the participating counties. Some conditionality on the use of the funds might be imposed, e.g., to ensure a certain amount of village-level infrastructure, to give villages or the community allocative decision-making power for part of the infrastructure funds, to ensure compliance with technical and other standards, to require sufficient budgets for operation and maintenance of infrastructure, etc., but this should be kept to a minimum. If deemed necessary, the incentive from increased allocative power as well as the timeliness of all funds might be further enhanced through the total amount provided. Given this example, the use of funding flexibility would require the monitoring and evaluation component to have a strong activity focusing on infrastructural services. The M&E of public services would be achieved through the strengthening of the capacity of the existing monitoring/data-collection system; through the commissioning of independent data collection; and through the commissioning of evaluations. For improved infrastructure investments, a spatial (based on Geographic Information System) planning tool would increase transparency of the geographic location of each individual infrastructure, for planning as well as for monitoring purposes. Monitoring, supervision, and checking require full information sharing and strong collaboration among the various technical line agencies. Participatory monitoring would be facilitated by the publication of the annual budget allocation for the infrastructure investments and the operation and maintenance budgets. However, reducing the scope of the services to be provided and enhanced under the pilot would limit the effectiveness of the reforms. Narrowing the scope of the pilot 73 was primarily for illustrative purposes.85 The narrowing of the scope would limit the pilot's ability to be fully responsive to the demands and priorities of the local population. Moreover, excluding parts of local public service provision would limit the ability to address systemic reforms, reduce transparency of local government actions since part of their activities would be excluded, and thereby reduce effectiveness. Using the example of the infrastructure-only focus, this would, first, not help to overcome the efficiency- lowering division between capital and recurrent investments, which leads to inefficiencies in local resource allocations. For example, counties or local PSUs might have a greater need for, say, operational expenses than for capital investments. Second, and at least as disadvantageous, it would severely limit the ability to leverage the funds for achieving improved service performance, in particular for the crucial social sectors where infrastructure investments play only a small role. Consequently, it clearly would be desirable to make the scope as broad as possible and to include all the types of rural public services that are provided in any given county. Only such a broad approach could fully ensure a fundamental shift in the county government's behavior to become a service provider. A broad focus would require close cooperation between the NDRC and the MOF, with strong support from the agencies responsible for personnel issues as well as other line departments. Such a broad scope scenario does not necessarily mean that improvements in all sectors would be included in the pilot's objective. However, it would require county government planning and budget processes that are transparent in all aspects/sectors. Box 7-1: Performance-based municipal infrastructure grants in South Africa In South Africa, the 284 municipal governments of the country are responsible for providing basic water, sanitation, and road infrastructure. The municipalities around the leading urban centers of the country finance their local services through their own generated revenues, but the rural municipalities that do not have similar revenue bases are overwhelmingly dependent on transfers from higher tiers. South Africa has consolidated the fragmented and ad hoc system of multiple channels of transfers to municipalities into a streamlined system of transparent and predictable transfers. By fiscal year 2006/07 the great majority of South African municipalities were receiving the following three types of transfers as per a Division of Revenue Act: (i) an unconditional Equitable Share and related transfers that meet the recurrent needs of local governments; (ii) conditional transfers to meet municipal investment needs, including for water, power, sports and recreation, and integrated rural development, now being consolidated into a Municipal Infrastructure Grant; and (iii) a set of conditional transfers for municipal capacity- building and restructuring. Much of the current focus is on fiscal monitoring--that the funds authorized for each municipality are indeed released by national and provincial departments in a timely fashion, and that the municipalities are properly informed about what they are supposed to receive and when. 85 Nevertheless, the approach could be effectively implemented by depending largely on the leadership of NDRC only. If the institutional barriers to the extremely close cooperation with the MOF required for a comprehensive reform are difficult to overcome, such an approach might be considered as a second-best alternative. Performance, in such a narrower focus, might be measured in terms of number and quality of rural schools, health centers, roads, water supply, etc. and, most importantly, the services that are provided through these investments to citizens. At the core of the reforms supported through such a pilot could be: (i) thorough planning of construction, renovations, and operation and maintenance of all public assets in a county, detailed by type of infrastructure, responsible level of government (county, township, village, community), and agency, etc. (ii) a sufficient budget allocation to ensure infrastructural services. 74 Municipalities are required to use various reporting formats to self-report on the funds they receive and how these are allocated and spent on different activities and investments. The bulk of grant monitoring activity currently deals with issues like what spreadsheets and software might be most useful for such reporting, and how municipalities with a low capacity in handling these formats might be assisted by program managers at the provincial level. The Municipal Infrastructure capital grants, for example, are being disbursed largely on the basis of past spending levels self-reported by municipalities, and validated through various reports on the assets being created. However, the South African National Treasury department is beginning to recognize the importance of using independent data on municipal public resource allocations and their impact on household and area development. In particular, the importance of household surveys administered by national-level agencies is gaining strength. A National Bureau of Statistics has been responsible for undertaking the census, a Labor Force Survey, and an annual October Household Survey. A review of these data-collection systems by the country's Council of Statistics concluded that although there was an abundance of national and provincial data, there was very little data at the local government level to develop representative estimates of local development. Following Cabinet acceptance of the Council of Statistics recommendation in 2004, the census was to be undertaken only every ten years in accordance with global practices, and the October Household Survey was to be replaced by the Community Survey, a household survey that provides data representation at the municipal level for program planning and policy formulation. The proposed scope of the survey is considerably larger than any existing household survey in the country, in the order of 170,000 dwelling units nationally drawn from 17,000 primary sampling units, with a probability-proportionate-to-size sampling frame to capture the spread of the various category municipalities. The "governance frontier" of the country has been recognized by the National Treasury department as one where objective indicators of municipal performance should be developed, and the information should be made publicly available so that citizens, administrators, and elected officials can bench-mark municipalities, learn from successful experiences, and monitor progress in lagging areas. Source: Girishankar et al. (2006). 75 REFERENCES Banerjee, A., R. Banerji, E. Duflo, R. Glennerster, and S. Khemani (2006). Can Information Campaigns Spark Local Participation and Improve Outcomes?, Working Paper No. 3967, Development Research Group, The World Bank, Washington DC. Barrera-Osorio, F. (2006). The Impact of Private Provision of Public Education: Empirical Evidence from Bogotá's Concession Schools, World Bank Policy Research Working Paper 4121, February. Bird, R. (2001). "User Charges in Local Government Finance," in Richard Stren and Maria Emilia Freire, eds., The Challenge of Urban Government, Washington: World Bank Institute, pp. 171-82. Bjorkman, M. and J. Svensson (2006). Power to the People: Evidence from a Randomized Experiment of a Community Based Monitoring Project in Uganda, Working Paper, IIES, Stockholm University. DRC (2006). China's Rural Public Service Reform and Development, Development Research Center, Report for the Ford Foundation, May 28. Fewsmith, J. (2006). Exercising the Power of the Purse? China Leadership Monitor, No. 19, Fall. Gale, F. (2002). China's Statistics: Are They Reliable? In USDA/ERS, China's Food and Agriculture: Issues for the 21st Century, AIB-775 ERS/USDA. General Office of the CPC Committee and General Office of State Council (2004). Guidelines on Improving the System of Openness of Village Affairs and Democratic Management, June 22. Girishankar, N., D. DeGroot, and T. V. Pillay (2006). Measuring Intergovernmental Fiscal Performance in South Africa: Issues in Municipal Grant Monitoring, Africa Region Working Paper Series No. 98, The World Bank, Washington, DC. Grout, P.A. and M. Stevens (2003). The Assessment: Financing and Managing Public Services. Oxford Review of Economic Policy, 19(2): pp. 215-34. Guggenheim, S., T. Wiranto, Y. Prasta, and S. Wong (2004). A Case Study from Reducing Poverty, Sustaining Growth: What Works, What Doesn't, and Why, A Global Exchange for Scaling Up Success, Scaling Up Poverty Reduction: A Global Learning Process and Conference, Shanghai. Holz, C. (2005). The Institutional Arrangements for the Production of Statistics, Statistics Working Paper, China Governance Project, OECD. Institute of Development Studies (1998). Participatory Monitoring & Evaluation: Learning from Change, IDS Policy Briefing Issue 12. Jin Renqing (2007). Report on the Implementation of the Central and Local Budgets for 2006 and on the Draft Central and Local Budgets for 2007, March 5, at http://www.10thnpc.org.cn/english/government/203405.htm. 76 Khemani, S. (2006). Local Government Accountability for Health Service Delivery in Nigeria, Journal of African Economies, 15(2): pp. 285-312. Lewin Group, Inc. (2002). Community Participation Can Improve America's Public Health Systems, prepared for the W.K. Kellogg Foundation, April. LGOPA (2003). An Overview of the Development-oriented Poverty Reduction Program for Rural China. Li Ping and Xu Hongcai (eds.) (2006). China: Intergovernmental Fiscal Relations. Beijing: China Public Finance Economics Press. Mackay, K. (2006). Institutionalization of Monitoring and Evaluation Systems to Improve Public Sector Management, ECD Working Paper Series 15, World Bank. Martinez-Vazquez, J. and R. Searle (eds.) (2006). Fiscal Equalization: Challenges in the Design of Intergovernmental Transfers. New York: Springer. Ministry of Education. Education Statistical Yearbook of China. Ministry of Education. China: Educational Finance Statistical Yearbook. Ministry of Finance. Budget Department (2004). Fiscal Capacity Transfers from the Central to Local Governments, Jingji yanjiu cankao (References on Economic Research), No.14. Muramatsu, Michio, Farrukh Iqbal, and Ikuo Kume (2001). Local Government Development in Post-war Japan. New York: Oxford University Press for The World Bank. NDRC (2004). A Statistical Study on Public Service Units, National Development and Reform Commission PSU Team, Internal Report presented at a workshop, June. OECD (2005). Governance in China, China in the Global Economy. Paul, Samuel (2002). Holding the State to Account: Citizen Monitoring in Action. Bangalore: Books for Change. Pi Shuchu and Li Pengxiang (2006). Who Shall Pay for Basic Education? A Survey of the Reform of Compulsory Education in Jianli County, Wuhan, Hubei Province, Xinhua Net, July 6. Plummer, J. and J.G. Taylor (eds.) (2004).Community Participation in China: Issues and Processes for Capacity Building. London: Earthscan. Qiang, Christine Zhen-Wei (2007). China's Information Revolution: Managing the Economic and Social Transformation, The World Bank, Washington, D.C. Ravallion, M. and S. Chen (2004). China's (Uneven) Progress Against Poverty, World Bank Policy Research Working Paper No. 3408, The World Bank, September. Rodden, J. (2003). Fiscal Decentralization and Challenge of Hard Budget Constraints (MIT Press). Sarraf, F. (2005). Integration of Recurrent and Capital "Development" Budgets: Issues, Problems, Country Experiences, and the Way Forward, The World Bank, Public Expenditure Working Group, July. 77 SDPC (2002). Feasibility Study Guideline for Investment Projects, State Development Planning Commission, China Power Publishing Press. Stanton, T. (2004). Reform of Public Service Units in China: Are There Lessons from the United States? Paper delivered at the Seminar on Public Service Delivery and the Role of Government: Experience of OECD Countries and Relevance for China, The World Bank and the NDRC, Beijing, March 23-24. State Council (1998). Tentative Regulation on Public Service Units Registration, Decree No. 252, October 25. State Council (2004). Reform of the Investment System, Guofa Paper No. 20. Stromberg, D. (2004). Radio's Impact on Public Spending, Quarterly Journal of Economics, 119(1): pp. 189-221. Tendler, J. (ed.) (1997). Good Government in the Tropics. Baltimore: Johns Hopkins University Press. Tsai, Lily L. (2004). The Struggle for Village Public Goods Provision: Informal Institutions of Accountability in Rural China, Paper presented at the Conference on Grassroots Political Reform in China, October 29-31, 2004. UNDP (2005). Development with Equity, China Human Development Report 2005 Wang Rong (2006). The Financing and Provision of Rural Education, Presentation at WBI Haikou Workshop on Rural Services, August. Wang, S., Sun, Q., Zhao, Y., et al. (2006). Public Health: A Case Study of Two Chinese Counties, Background Report for China Rural Health Report, The World Bank. Wen Jiabao (2007). Report on the Work of the Government, March 5, at http://news.xinhuanet.com/english/2007-03/16/content_5857166.htm. West, L.A. (1997). Provision of Public Services in Rural PRC. Chapter 6 in: Christine Wong (ed.), Financing Local Government in the People's Republic of China. Hong Kong. Oxford University Press. Wong, C., C. Heady, and W. Woo (1995). Fiscal Management and Economic Reform in the People's Republic of China. Hong Kong: Oxford University Press. Wong, C.P.W. (ed.) (1997). Financing Local Government in the People's Republic of China. Hong Kong: Oxford University Press. World Bank (2002). China: National Development and Subnational Finance: A Review of Provincial Expenditures, Poverty Reduction and Economic Management Unit, East Asia and Pacific Region, Washington DC. World Bank (2003). Case Study 2 - Porto Alegre, Brazil: Participatory Approaches in Budgeting and Public Expenditure Management, Social Development Notes, Note No. 71, March. World Bank (2004a). Making Services Work for the Poor, World Development Report. World Bank (2004b). Citizens, Politicians and Providers: The Latin American Experience with Service Delivery Reform, Draft Report. 78 World Bank (2004c). The Governance of Service Delivery in ECA, A Regional Study, ECAVP, June 30. World Bank (2004d). Sierra Leone ­ Institutional Reform and Capacity Building Project, Project Appraisal Document. World Bank (2005a). China: Deepening Public Service Unit Reform to Improve Service Delivery, Poverty Reduction and Economic Management Unit, East Asia and Pacific Region, http://www.worldbank.org.cn/english/content/psu_en1.pdf World Bank (2005b). East Asia Decentralizes: Making Local Government Work, The World Bank, Washington, DC. World Bank (2005c). Kenya ­ Institutional Reform and Capacity Building Technical Assistance Project, Project Appraisal Document. World Bank (2006a). Development and the Next Generation, World Development Report 2007. World Bank (2006b). Bangladesh ­ Local Governance Support Project, Project Appraisal Document. World Bank (2007a). China: Rural Public Finance for a Harmonious Society, The World Bank, unpublished. World Bank (2007b). Reforming Subnational Finance: Lessons from Northeast China, The World Bank, unpublished draft. World Bank (2007c). Meeting China's Rural Health Challenges, China Rural Health Study Publications and Reports, http://go.worldbank.org/QAVXF4NUO0 World Bank (2007d). China's Evolving Poverty Reduction Agenda: An Assessment of Poverty and Inequality in China, unpublished draft. Xu, Lin and Mei Wang (2007). A Monitoring and Evaluation Framework for China's 5- Year Plans, Report on IDF Project "Monitoring and Evaluation of Development Process," Forthcoming. Xu Ping (2006). Cultural Construction of New Socialist Countryside, Scientific Socialism, No. 1. Zhang Minyu (2006). Community Led Public Sanitation Mobilization, Plan China Newsletter, No. 1, March Zhao Yang et al. (2006) Final report; TCC4 activities supported under the World Bank ­ financed TCC4, report submitted to MOF Agriculture Department, unpublished, October 18, 2005 Zhou Jian (2006). Presentation by Zhou Jian, the Director of the Compulsory Education Guarantee Scheme, Ministry of Education, at a National Academy of Educational Administration-World Bank Workshop, November 6-8. 79 Background Notes and Reports in the Files: Aide-Memoire, Chongqing mission from July 12-20, 2006, August 2006, Unpublished. Aide-Memoire, Ganzhou City / Jiangxi mission from November 7-20, 2006, November 2006, Unpublished. Huang Hanquan (2007). Research on the Role of the NDRC on Providing Public Services in the Rural Society, Infrastructure and Agriculture Construction, Report for The World Bank, Unpublished, February. Khemani, Stuti (2007). Using Intergovernmental Transfers to Promote Equity in Service Delivery: Lessons from International Experience, Background Paper, China: Rural Public Service Delivery, February 2. Lin Wanlong and Sun Cuiqing (2006a). Delivery of Rural Public Services, Report for The World Bank, Unpublished, October. Lin Wanlong and Sun Cuiqing (2006b). List of Pilot Reforms, Report for The World Bank, Unpublished, October. Liu Mingxing (2007). Assessment of Local Incentives and M&E for Provision of Compulsory Education in Rural China, Report for The World Bank, Unpublished, January. Zhang Changchu (2006). NDRC's Role in Public Services for Rural Development, Unpublished report submitted to The World Bank, Beijing, November/December 2006. 80 Annex 1: Assignment of Responsibility in the Health Sector Relevant institutions and Roles and responsibilities agencies Central Ministry of Health, CDC, ·Broad policy and guidelines National Population and ·Financing of central institutions Family Planning ·Some matching grants for investments Commission at lower levels ·Technical assistance and supervision of lower-level institutions (limited) Province Provincial Health Bureau, ·Develop and adapt central government CDC, MCH Center, Family policy and guidelines Planning Station, Disease- ·Financing and operation of provincial specific institutes, Center public health institutions for Health Inspection and ·Some matching grants for investments Supervision, Provincial at lower levels hospital ·Technical assistance and supervision of lower-level institutions (limited) Prefecture City Health Bureau, CDC, ·Financing and operation of city-level / City MCH Center, Family public health institutions Planning Station, Disease- ·Some matching grants for investments specific institutes, Center at lower levels for Health Inspection and ·Technical assistance and supervision of Supervision, City hospitals lower-level institutions (limited) County County Health Bureau, ·Financing and operation of city-level CDC, MCH Center, Family public health institutions Planning Station, Disease- ·Some matching grants for investments specific institutes, Center at lower levels for Health Inspection and ·Technical assistance and supervision of Supervision, County lower-level institutions (limited) hospitals Township Township administration, ·Supporting operations of township THC, Family Planning health center and village clinics, which Station should perform key public health functions Village Village doctors, Family ·Provision of public health services and planning cadres, Informants implementation of collective public health activities according to mandates or contracts 81 harbor balance traffic farmland Governments and secondary markets, Local security, demand airports, by networks, projects education local and government public universities, basic Financed Irrigation construction Regional Manage supply Highways, construction Local administration Local control Local and 2005) control (in and system agriculture major and river local forest Financed cases safety, crime, research 30% extension major Assignments food crime Jointly prevention natural central, Actual Agricultural technology building, disaster alleviation, projects, protection 70% Cross-regional economic 82 the which anti- Financing on ensure and Education Joint Central areas, regions projects impact aviation universities of or by cross-drainage drugs, major poor and society market, safety civil Local Government environmental in in expenditures and and Ministry grain grain Financed significant government education, projects, a the under Ecological protection assistance Cross-regional area have economy Control national Railways Central administration Anti-terrorism, smuggling Higher under and Assignments Security, Item Agriculture Infrastructure Grain Comprehensive Agriculture Development Transportation Administration Public Procurator, Courts Education Current 2: to local jointly central Annex Items be financed by and gov'ts Governments administration administration administration administration administration institutions, work Local construction, departments and and and and and by science research cultural Financed Local infrastructure popular Local Management Management Management Management Management 2005) (in subsidizes local Financed local 30% 10% Assignments Jointly governments deficits, central, of Actual 70% Central 90% governments 83 and provides provides provides Central by institutions, Foundation, departments, central cultural in Government subsidies subsidies subsidies Science projects for research cultural regions government government government Financed (2006). Central National scientific Central assistance development western Central earmarked Central earmarked Central earmarked Xu and Li Care Relief Security Social from Item Science Culture Health Unemployment Insurance Medical Insurance Social and Welfare Social Adapted to local Items be financed by gov'ts Source: Annex 3: Tables on Education Financing Annex 3, Table 1: Education Earmarked Funds Provided by Central Finance, 2003 Item RMB m 1. Earmarked Funds for Consolidation of Primary and Middle Schools 1000 2. Second Phase "Compulsory Education Project" Funds 800 3. Boarding School Renovation Funds 600 4. Subsidies to Exemplary Vocational Schools 250 5. Free Textbook Subsidies for Fall Quarter 230 6. Funds for Improving Education Conditions of Schools in Border Areas 200 7. Subsidies for Improving Primary and Secondary School Conditions in 183 Extremely Poor Areas (Education Fee Surcharges) 8. Funds for Education Disaster Alleviation 152 9. Subsidies for Local Education Infrastructure (High Schools) 112 10. National Compulsory Education Assistance Funds 100 11. Second Phase "Compulsory Education Project" Reserved Funds 100 12. Free Textbook Subsidies for Spring Quarter 93 13. Free Textbook Subsidies for Xinjiang 77 14. Subsidies for Local Education Infrastructure (Normal Schools) 55 15. Subsidies for Minority Textbook Publishing Deficit 30 16. Subsidies for Local Education Infrastructure (Special Difficulties) 22 17. Subsidiary Funds for Education of Disabled Students 15 18. Other Earmarked Funds 14 19. Subsidies for Local Education Infrastructure (Special Education) 10 20. Bonus Funds for Reduction of Illiteracy 8 21. Funds for Tibet Inland School 7 22. Subsidies for Local Education Infrastructure (Xinjiang Production and 5 Construction Corps) Total 4062 Source: Wang Rong (2006). Annex 3, Table 2: Central-Local Division of Recurrent Budgetary Education Expenditures 1998 1999 2000 2001 2002 2003 2004 2005 Total (bn RMB) 133.8 152.3 176.5 210.8 264.5 293.7 336.6 397.5 Central 12.3 12.7 14.1 17.2 21.0 24.0 22.0 24.5 Sub-national 121.5 139.5 162.4 193.6 243.5 269.7 314.6 373.0 Share of education in total budget National (%) 12.0 11.3 10.9 11.0 11.9 11.8 11.7 11.6 Central (%) 3.9 3.0 2.5 3.0 3.1 3.2 2.8 2.8 Source: Calculated from CSY (1999-2006); CFY (2005); PFSY (1998-2004). 84 Annex 3, Table 3a: Government appropriations for rural basic education (MOF data) 2002 2003 2004 Junior Secondary Schools (bn 34.14 37.64 46.83 RMB) Share of total 79.9% 80.3% 81.5% Primary Schools (bn RMB) 71.01 76.64 92.53 Share of total 84.6% 85.5% 86.4% Source: CSY (2004-2006). Annex 3, Table 3b: Government appropriations for rural basic education (MOE data) 2002 2003 2004 Junior Secondary Schools (bn 30.7 34.2 42.3 RMB) Share of total 72.3% 73.5% 74.1% Primary Schools (bn RMB) 63.4 69.4 83.1 Share of total 75.9% 77.8% 78.1% Source: CEFY (2003-2005). 85 Annex 4: Challenge in M&E for performance-based transfers The challenge for countries like South Africa, Chile, and Brazil that are pursuing performance-based transfers to improve public services lies in setting up data collection systems that reliably and credibly measure household access to services and service delivery performance. Such data could then be used to target resources to areas where there is the greatest need, and subsequently to monitor performance improvements. Chile, for example, has invested considerably in an M&E system managed by the powerful Ministry of Finance. Since 2001, Chile has been undertaking rigorous impact evaluations of specific spending programs by contracting out the primary data collection and analysis to independent research organizations and academics. In countries that systematically collect detailed household-level data on services access on a sample basis, the M&E systems are typically delegated to semi-independent national agencies such as a National Statistics Office or a Census Bureau. In countries with a "socialist" tradition, where agencies analogous to the NDRC exist--such as the Planning Commissions in India and Nigeria--these agencies also have substantial authority to engage in primary data collection to evaluate specific spending initiatives, often by contracting out to independent civic organizations, research outfits, or academia. In the United States various government departments have offices that are mandated to regularly collect household data on a sample basis. Since 1940 the Bureau of Labor Statistics has been conducting a monthly sample survey, the Current Population Survey, to measure unemployment, and the Survey of Income and Program Participation; the US Department of Health and Human Services conducts an annual survey on drug use and health issues, the National Survey on Drug Use and Health and the National Health Interview Survey; the Bureau of Justice Statistics undertakes an annual household survey on crime, the National Crime Victimization Survey. In addition to the decennial census, the Census Bureau of the United States also undertakes an annual Consumer Expenditure Survey. In India, detailed household surveys on living conditions and access to services are undertaken regularly on a sample basis by the National Sample Survey Organization under the Ministry of Statistics and Program Implementation. This database is regularly used to calculate various indicators of socio-economic well-being across regions in India that become the bases for determining the distribution of different types of transfers and government expenditure programs. In order to target programs for maternal and child health to particular lagging districts, the Ministry of Health and Family Welfare commissions the International Institute of Population Sciences to conduct annual household surveys to measure reproductive and child health at the district-level. Apart from these national-level agency data collection efforts through rigorous sampling and survey techniques, most countries have systems of reporting by lower-level agencies and providers on detailed aspects of public services. There are substantial problems of reliability and credibility in this type of reporting data because of local incentive problems in monitoring their own behavior. Public Expenditure Tracking Surveys in various countries have demonstrated that data reported by local governments cannot be trusted. Source: Khemani (2006). 86 Annex 5: Reforming Rural Production and Cultural Services: The Hubei Approach Under the leadership of the CPC Hubei Provincial Committee, Hubei Province has piloted a decisive reform of rural production and cultural services. The pilot started in seven counties in 2003 and since 2005 has been extended to all counties. An initial evaluation conducted by the provincial authorities in mid-2006 suggests desirable results. The central idea of the reform lies in the slogan "financing services, not employees" (yangshi bu yangren). Key components of the Hubei approach include: A. Formulating a catalogue of services to be financed ­ and purchased ­ by the government. A provincial guidance catalogue includes 74 items, based on which counties are granted the autonomy to formulate their own catalogues. Significantly, each participating level of government is required to ensure a substantial budget increase for rural production and cultural services. B. Abolishing all township PSUs (except for the "fiscal institute") and terminating the employment relationship with all their employees. This is made possible by well- defined compensation schemes for formal PSU employees. In particular, significant public financial resources are made available by the provincial budget to enable all formal PSU employees to join the enterprise pension insurance scheme. Extra compensation is offered (one month salary for one year of service) to encourage formal PSU employees to give up their government employee "status." "Self- financed" and other informal employees of PSUs are simply laid off without any significant compensation. C. Reorganizing service providers. The government is issuing qualification certificates to whoever has passed a test on technical capacity and wishes to compete in the markets for rural production and cultural services, including those who have never been PSU employees. The government also encourages the creation of "service centers," registered as "minban (privately-run) non-enterprise units." Their organizational features seem to be still evolving. In some counties (such as Yuan An County), they resemble more an association than a "unit," as individual service providers join the center as "members" rather than as "employees." The service centers typically receive the township governments' permission to use the old PSUs' state-owned assets (offices, equipment, etc.) for free or against a rental fee. D. Packaging services for public offerings. Services in the catalogue of government purchasing are packaged as contracts for public offerings by three entities: the village committee, the township government, or the county government, depending on the scope of the beneficiaries of the service in question. The county budget takes all the bills, so that the division of contracts among the three levels is equivalent to a division of the county budget for these services. In the case of Yuan An, which has 7 townships and 110 villages, there were 681 contracts offered in 2005 and 311 in 2006. E. Offering the contracts through a public process. On a publicly announced date, the three entities offer all the contracts at a publicly announced location. All the service centers and individual service providers that have a qualification certificate can join the competition and can sign a contract with one of the three entities. Service centers and individual providers of one township can compete for contracts offered by any 87 other township in the county, and former PSU employees have to compete equally with others. In Yuan An County, 35 percent of the contracts offered in 2006 (50 percent in 2005) were granted to providers who were not former PSU staff. Presumably, some services that used to be delivered by PSUs are not included in the catalogue and therefore are not covered by the contracts. These services are regarded as commercial in nature and left to the private market, delivered by private providers that charge farmers to recover costs. F. Monitoring and evaluating performance through carefully designed schemes, with significant local variations. Yuan An County, for example, has established a 100- point evaluation scheme for four entities: the service center, of which the service provider is a member, the village committee, the township government, and the relevant line bureau of the county government. Different weights are assigned to the four entities in different sectors. In the case of agricultural extension services, for example, the weights of the four entities are 15, 50, 15, and 20, respectively. Farmers' voices are reflected in the evaluation of the service center, which collects farmers' signatures for services delivered, and in the village committee. G. Linking performance with pay. A service provider can only receive less than 50 percent of the contracted payment before the M&E. The remainder is paid at a discount in line with the total score s/he receives in the M&E process. For example, a score of 90 will mean that the service provider is entitled to only 90 percent of his/her contracted payment. Anyone who receives a score below 60 will be disqualified from the competition in the following year. Source: Office of the Leading Group for Township Comprehensive Reform of Hubei Province, Compilation of Policies and Pilot Materials. Other materials provided by Hubei Provincial Development and Reform Commission, Yuan An County, and Yang Ping Township of Yuan An County. 88 Annex 6: Private Management of Public Schools: Examples from Latin America Governments in Latin America have sought innovative ways to overcome the substantial difficulties in reforming bureaucracies (central and local) as a means of improving service delivery outcomes. Many of these alternatives involve a combination of contracting services to private firms and non-government organizations (NGOs) and relying on alternative bureaucratic structures that bypass more traditional structures. Two of the many well-documented and successful initiatives include: The Bogotá School Concessions Program. The municipal education department of Bogota constructs new, state-of-the-art schools in low-income areas. School management is tendered in a public procurement process; bidders are private educational institutions that are evaluated based on their proposed management plans. Concessionaries receive remuneration from the government department on a per-student basis. Although the contract establishes clear standards that must be upheld, the concessionaries have full autonomy over school management and are evaluated solely on results. Concessionaries have already produced striking results in terms of management improvements: they allocate on average 55 percent of the per capita remuneration to human resources, well below the 90 percent allocation by public schools. Fe y Alegría Schools operate in 14 Latin American countries covering approximately one million students. Evaluation of the program's operations in nine countries indicates that schools integrated into Fe y Alegría outperform traditional public schools, with lower repetitions and drop-out rates, and higher overall rates of progress in school and higher overall retention rates on average. Critical to Fe y Alegría's success are its parent- centered strategies based on an open school environment that promotes effective parental involvement in the learning process, community involvement in the management of the schools, and student-centered strategies that depend on an interactive teaching method and recognition of a variety of learning styles. The dynamic at play in this case is strongly associated with the altruistic and "empowerment" oriented approach that characterizes Fe y Alegría. In other words, the effects of contracting out appear to be closely related to the nature of the contractor. Source: Barrera-Osorio (2006); World Bank (2004b). 89 Annex 7: Institutionalizing Participatory Monitoring and Evaluation 1) Establish the monitoring and evaluation group at the community level. 2) Design and manage the Participatory M&E process jointly by service responsive government officials and community monitoring group members. Different aspects should be covered by Participatory M&E, such as input, service quality, process, performance of PSU, and attitudes of community members to services. 3) Set up clear responsibilities for government officials and community monitoring group members in the Participatory M&E process. Government should share the development objectives of public services, such as enrollment rates, attendant rates for basic education, and coverage range of the NCMS, with communities and citizens, to publicize the technical standards of public service operations, such as indicators of village school closures, norms of village clinic equipment, and requirements of village doctors and in the community. A community monitoring group is mainly responsible for collecting information. 4) Hire an external community facilitator working at the community level. The main responsibility of the external community facilitator is to: a) provide training to local people in various areas, including in participatory skills needed for participatory monitoring; b) provide government with the community feedback on public services; and c) provide guidance to the community as needed. 5) Set up a hotline in the corresponding government sector to handle complaint feedback from the community. 6) Assign responsible government officials in different service sectors to document the findings of the Participatory M&E and the concerns and feedback from the community on public services to decision makers, then to be reflected in the planning, policies, and programs of public service provision. 90 Annex 8: CDD Pilots for Rural Poverty Reduction in China LGOPA in collaboration with the World Bank is undertaking a CDD pilot program in fifteen villages of four counties and provinces. It was launched in May 2006 and will be completed in October 2008. The program goal is to assess the suitability of CDD under Chinese field conditions as a means to provide poor communities with resources necessary to sustain their own development by strengthening their self-capacity to manage and organize their own development activities, including small-scale infrastructure, public services, environmental improvements, natural resource management, and community development funds for income generation. Two key phases are designed in the pilot operational processes, i.e., community preparation and implementation of development activities. Community preparation must be accomplished before participating villages can obtain access to program funds. In this phase, community facilitators are selected and trained, participating communities learn about the program and its objectives, natural village members elect a Project Management Committee, and its members select representatives to join a Project Decision-Making Group at the administrative village level. Following completion of community preparation, participating communities can initiate their development activities in three rounds by following the project cycle. The cycle for the small-scale infrastructure and public service component includes the following steps: 1) Identification of Priorities. The program operation cycle typically begins with one or more natural village meetings to identify local priorities and to decide on specific activities to propose for program funding. 2) Preparation of Proposals. The Project Management Committee prepares proposals for funding, with assistance from the village facilitator. 3) Preliminary Selection of proposals for funding on a competitive basis, with selections done by the Project Decision-Making Group representing all natural villages. 4) Screening and Review by county or township governmental agencies for cost estimations, technical feasibility, and acceptability under program rules. 5) Proposal Finalization. The Project Management Committee revises the proposal as warranted. A village meeting is called to confirm the acceptability of the proposal and to make arrangements for implementation. 6) Disbursement of Funds to a joint account managed by Project Management Committee members. This step begins the implementation process. 7) Monitoring is conducted by a county project management office, by community facilitators, and through a participatory inter-village monitoring arrangement. 8) Interim Review of financial management and technical implementation. Participating villages found to be in compliance with program rules will be eligible for the next phase of the cycle. Source: China CDD Pilot Project, Community Procedures Manual, May 2006. 91 Annex 9: International Experiences of CDD Approaches CDD application poverty reduction programs in Indonesia. The CDD approach has been widely applied in nationwide poverty alleviation programs. They have been supported by national and local government internal funding, donor grants, and World Bank loans. A prominent example, the Kecamatan Development Program (Guggenheim et al. 2004) began in 1998 at a time of tremendous political upheaval and financial crisis. Currently, the program is in its third phase, and it is expected to run until 2009. The program started in 28 villages on a pilot basis. Eight years and 65,000 facilitators later, the program is going national with full ownership by the government. This occurred only after consistent economic returns in the 30-40 percent range, strict discipline in penalizing poorly performing kecamaten (districts), and after numerous communities had won rights over their village leaders who tried to hijack funds. Since 1998, the Kecamatan Development Program and a parallel CDD-type Urban Poverty Program have covered some 41,000 rural and urban villages in Indonesia. They have increased community participation in governance and have found creative new ways to combine formal and informal anti-corruption programs. The programs have also been instrumental in delivering capacity-building and governance-support programs at the local level. The programs have been subjected to the most intensive independent audits. This public scrutiny contributes to making the Kecamatan Development Program and the Urban Poverty Program attractive choices for leading the country's national poverty reduction strategy. Under the new National Community Empowerment Program, all 70,000 villages in the country will be covered. Community participation in education in the Gambia. The government of the Gambia has made good progress since the early 1990s in narrowing the gender gap in enrollment at the lower basic level. This achievement can be attributed to a good base for promoting community participation in school development. The Gambian government piloted a model of Whole School Development that puts communities, in the shape of School Development Committees, center-stage in school development planning. Through awareness-raising, skill training, and confidence building, mothers and fathers both play an active role in the School Development Committees. Community participation in public health in the United States. Both policy-makers and most Americans often see public health primarily as a provider of clinical services for poor people, and public health's critical population-based responsibilities are not understood. With funding from the W.K. Kellogg and Robert Wood Johnson Foundations, 41 local communities in 14 states accepted the challenge to transform their public health systems by creating public health partnerships. Some partnerships successfully engaged a wide array of community sectors in public health work (Lewin Group 2002). wb16201 N:\CHINA\AAA-Rural Public Services\GreycoverRPS 2007 Board version.doc 12/14/2007 2:24:00 PM 92