Report No. 32341-CHA China Deepening Public Service Unit Reform to Improve Service Delivery July 12, 2005 Poverty Reduction and Economic Management Unit East Asia and Pacific Region Document of the World Bank TABLE OF CONTENTS ACKNOWLEDGMENTS ............................................................................................. ... PREFACE........................................................................................................................ 111 v EXECUTIVE SUMMARY .......................................................................................... vii Public Service Provision inChina ............................................................................. vii InternationalExperience InPublic Service Provision andReform............................. vii Service providers inOECD countries.................................................................... vii vm ... Towards a reform strategy for China's PSUs .............................................................. Public service reforminthe world ....................................................................... ix Improvefinances .................................................................................................... ix Redefinethe role of government ............................................................................. ix Reorganize the PSU sector....................................................................................... x Manage the impact on employees ........................................................................... Organize Reforms................................................................................................... xi Strengthen accountability......................................................................................... x xi INTRODUCTION........................................................................................................... 1. CHINA'S PUBLIC SERVICE UNITS AND THEIR REFORM: AN 1 What are PSUs andHow Do They Operate?................................................................ 1 How have PSUs beenReformed inthe Past?................................................................ 3 Overall strategy and classification............................................................................ 4 Increasedprivate participation ................................................................................. Pushing into the market ........................................................................................... 4 5 6 What are the Problems now?........................................................................................ Introduction of the notion of "not-for-profit" ........................................................... 8 Financing.,............................................................................................................... 8 12 What Reform Strategies have been Proposed?........................................................... Incentives .............................................................................................................. 15 The Cheng Siwei Study ......................................................................................... 15 The DRC study...................................................................................................... 15 The NDRC Study................................................................................................... The MOST Study................................................................................................... 16 16 REFORM...................................................................................................................... 2. INTERNATIONAL EXPERIENCEINPUBLIC SERVICE PROVISION AND ITS 19 19 Public Service Providers inOECD Countries............................................................. The Finance of Public Services inOECD Countries................................................... 20 Government bodies involvedinpublic service delivery., ........................................ 22 25 Non-profit organizations ........................................................................................ Private sector participation..................................................................................... 26 Public Service Reformin OECD Countries................................................................ 27 The case of New Zealand....................................................................................... 27 The case of the UnitedKingdom............................................................................ 28 i Public Service Reform inTransition Countries .......................................................... Financing.............................................................................................................. -30 29 31 Public service reform inLatin America andthe Caribbean......................................... Provision ............................................................................................................... 31 32 Establishment of compacts with non-government providers ................................... Progress incoverage of basic services.................................................................... 32 Empowerment of clients with choice...................................................................... 33 34 Lessons for China...................................................................................................... Client participation ................................................................................................ 34 3. TOWARDS A PSU REFORM STRATEGY.......................................................... 37 -37 Improve Finances ...................................................................................................... Redefine the role of government ............................................................................... 39 Reorganize the PSU sector......................................................................................... 40 42 Organize Reforms...................................................................................................... Strengthen accountability........................................................................................... 45 Manage the Impact on Employees ............................................................................ -46 REFERENCES.............................................................................................................. 49 ii ACKNOWLEDGEMENTS This study was undertaken by a core task team ledby ChunlinZhang (Task Team Leader) and including Christine P. Wong and Edward Mountfield. The final report i s written by Chunlin Zhang with specific advice from Bert Hofman. Substantive inputs to the report were provided by Allen Schick (international experience and direction of China's reform), Elsa Pilichowski (OECD experience), Nan Jia (experience of transition economies), Xiaoqing Yu and Yee Mun Sin (labor market and pension reform), Eduardo Velez Bustillo, Rong Wang, Yidan Wang, Halsey Beemer, and Jeffrey Waite (education sector case study), Magnus Lindelow and Ling Li (health sector case study), Achim Fock and Bingsheng Ke (agriculture extension case study). Xiuzhen Zhang provided excellent assistance to the team. Gloria Elmore and Michael Figueroa assisted inthe production of the report. The team benefited greatly from partnerships with the Organization for Economic Cooperation and Development (OECD) and the Department for International Development (DFID) of the UnitedKingdom. The study has been carried out under the guidance of Director General Fan Hengshan, Deputy Director-General Zhang Lina of the National Development and Reform Commission (NDRC). It also benefited from guidance given by Director General Dong Hong, Deputy Director-General Wang Qiang of the former State Council Office for Restructuring Economic Systems (SCORES) in its early stage. The team wishes to express sincere thanks to all members of the NDRC PSU Study Team for sharing their valuable insights as well as data and materials. The team i s also indebted to a large number of government officials and PSU managers in Zhejiang province, Yunnan province, Jiangsu province and Nanjing city, where the team was offered generous support in field .work. The team was warmly received in a number of PSUs located in Beijing as well. Earlier versions of this report were discussed intwo workshops held in Beijing in October 2003 and June 2004, where valuable comments were received from participants, including in particular, Lin Zhaomu, Xue Lan, Lu Wangshi, Gao Shiji, Ge Yanfeng, Tan Songhua, Lei Haichao, Hu Ruifa, Cai Renhua, Li Keshi, Wang Fenyu, Zhang Yalin and LuHoushan. The study has also benefitedfrom insightsof a group of experts broughtto an international seminar in Beijing in March 2004 by the OECD: among them, Miroslav Beblavy (Slovak Republic), Derek Gill (New Zealand), Mary Tetlow (United Kingdom), Ronald Van Oosteroom (Netherlands), Thomas Stanton (United States) and John Mayne (Canada). Insightfulcomments from peer reviewers Nick Manning, Anwar Shah and April Harding, as well as Sector Manager Barbara Nunberg, are appreciated. Essential guidance and advice were given by Deepak Bhattasali and Bert Hofman, Lead Economists for China. Homi Kharas, Chief Economist and Sector Director, East Asia and Pacific Region, Yukon Huang and David Dollar, Country Directors for China, provided overall guidance. iii PREFACE This report presents findings of a World Bank study on China's PSU reform, conducted during 2002-2004. The study was designed to develop an overall strategy for PSU reform in China, drawing on systematic analysis of the current situation and critical assessment of international experience with public services reform. Realizing the diversity and complexity of the PSU sector and the critical importance of sector- and region-specific constraints, this study concentrates on a range of important cross-cutting themes based on case studies of three key sectors; namely, education, health and agriculture extension services. It i s intended to contribute to the development of a PSU reform strategy by: (i)providing an alternative way of thinking that may help policymakers think through various issues involved in the development of PSU reform strategy; and (ii)recommending key guiding principles and direction of reform actions. Constrained by the complexity of the issue and availability of data, the study does not address practical implementation issues, except for a few. Further work i s needed to translate the recommended overall strategy into implementable action plans tailored to specific circumstances of sectors andregions. The report consists of three chapters. Chapter 1provides an introductory account of what PSUs are, how they operate, how they have been reformed in the past two decades, as well as a diagnostic analysis of existing problems of the existing system. Relevant international experience with public service delivery and reform i s reviewed in Chapter 2. Key recommendations with regard to an overall reform strategy are presented inChapter 3. V EXECUTIVESUMMARY PUBLIC SERVICE PROVISIONIN CHINA 1. China has more than one million public service units (PSUs), or shiye danwei, with a labor force of around 30 million. Most PSUs were created as public service providers. Non-state involvement in service delivery remains limited in China today. PSU performance i s therefore crucial for improving service delivery in the public sector to a level commensurate with China's stated goals of a "well-off (xiaokang)" society and "people-centered" development. 2. Past reform efforts, aiming to "push PSUs into the market," have achieved much, but at the same time introduced undesirable incentives in services delivery. Nearly half of PSUs' funding i s raised through charging fees, which often cross-subsidize the public service delivery of the unit, but also allow for bonuses and welfare for staff on top of formal salaries. This gave a strong incentive for PSUs and their supervisory departments to distort the market in which PSUs are operating. With the rise in income inequality, reliance on user charges for financing service delivery i s increasingly becoming a barrier to access for the poor. In addition, greater autonomy in revenue generation often was not accompanied by better performance management and stronger financial accountability. The efficiency of PSUs also suffered from overstaffing. In recent years the government has made very large additional injections of funds. This, however, does not appear to have produced a corresponding decline infees, with a few exceptions. Many signs point to the additional resources being increasingly captured by personnel expenditure. INTERNATIONAL PROVISIONANDREFORM EXPERIENCEIN SERVICE PUBLIC Service providers inOECD countries 3. Organizationalforms. Service providers in OECD countries show a wide variety in organizational forms, including different budgetary relations, funding sources and govemance. In contrast to China, there is considerable involvement of non-state organizations inthe production of public services, even if financed from the budget. And again unlike China, there are few public agencies that are involved in predominantly commercial activities. Within the public sector, the legal form that a public service provider can take i s more diverse than the all-purpose model in China. Service providers differ widely in terms o f autonomy and accountability. Traditional ministries and state owned enterprises define two extremes of the spectrum. Inbetween lies a diverse range of public entities, of which "departmental agencies" and "public law administrations (PLAs)" are two of the larger groups. Departmental agencies can be considered as "delegated" service providers. They are direct subsidiaries of ministries and function under public law or general administrative processes applying to all ministries. Public vii viii Deepening Public Service Unit Reform to Improve Service Delivery Law Administrations are a diverse group. Within any one country, a wide array of governance regimes and legal forms may be used to undertake a wide variety of functions. Across countries the legal status, control regimes, powers, size, functions, internal governance and administrative cultures vary even more. PLAs can be seen as "devolved" service providers. Their main difference with ministries lies in their separate legal entities from the state and their different governance structure. 4. NPOs. Alongside the public sector and for-profit private sector, there can be a "third sector": Non-Profit Organizations or "0s. A NPO i s an organization that is barred from distributingits net earnings, if any, to individuals who exercise control over it (known as the "non-distribution constraint"). The value of the NPO as an institution i s primarily about trust and motivation. By institutionalizing a weak profit motive, "0s help buildup trust of consumers and donors toward service providers. Such trust reduces transaction costs and makes the market work for some transactions that would otherwise be unattainable. Interms of financing, the NPO represents a complement or alternative to the government and the for-profit sector when it finances services with donations. OECD countries differ greatly in the extent to which various services are provided by the " 0 sector. Public service reform inthe world 5. OECD countries. In recent decades, public sector organization and management has undergone major reforms in several OECD countries, particularly in New Zealand, United Kingdom, Netherlands and Sweden. The New Zealand model went furthest in introducing market-like mechanisms into the public sector, including the introduction of separation of purchaser and provider, establishment of a public services commission to manage performance contracts, and major reforms inbudget management and accounting to better reflect true costs of policies. 6. Transition countries. Service delivery has suffered greatly from financial difficulties in the 1990s in most transition countries. In some countries, the combination of inadequate funding, poorly designed user charging and weak institutions for regulation, transparency and accountability have damaged public service ethics and led to an increase in administrative corruption in social service delivery. Many transition countries have embarked on privatization programs which have included large service- providing entities. Some countries have attempted to establish "functional review" mechanisms in order to redefine the boundaries of the state by abolishing, rationalizing, privatizing or transferring functions of government, and by restructuring and reforming remaining functions. Overall evidence i s mixed as to how such processes work in practice. However, in some particular cases, such as primary health care, there have been successes. 7. Latin America and the Caribbean countries. The last two decades have witnessed radical reform in public service delivery in most countries in the Latin America and Caribbean region. Pronouncedimprovement has been made incoverage of basic services with increased public spending. Establishing compacts with agencies or actors outside the mainstream state apparatus has produced good results overall, at least inthe short run. Executive Summary ix Giving individuals choice in selecting service providers has, inthe limitedareas inwhich it has been implemented, generated powerful pressures for results on front-line providers. Client participation in the management and operation of services has also shown good results in the case of services that allow a relatively small production scale. However, large inequalities in access to services remain a feature in most countries in the region, and reforming the traditional state institutions remains a fundamental challenge. TOWARDSREFORMSTRATEGYFORCHINA'S PSUS A 8. 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, revamping public finance for public services, allowing for more non-state supply of public goods, improving accountability relationships within the public sector, and stepping up performance management and monitoring. Given the complexity and diversity o f PSUs, it i s crucial for such reforms to be designed and implemented with full consideration of sectoral and regional circumstances, as well as their impact on employees. Creation of a central, multi-ministerial task force to guide and monitor reforms may well be desirable. Redefine the role of government 9. While China has made great achievements in the transition to a socialist market economy, a reconsideration of the role of the state in the service sector i s warranted. Many of the activities performed by the state could be considered commercial production of private goods or services, which can best be left to the market. The concept of "public benefit relevance (gongyi Xing)," which i s often considered the defining criterion for government involvement, i s only a necessary, but not a sufficient condition for government intervention. Only in the case of market failure should government intervention be considered. Improvefinances 10. China's government spending i s not low by East Asian standards. However, services considered as key to achieving national goals are underfunded, especially in the poorer regions. This calls for better allocation o f public resources over services, and across the country. In addition to the transfer of some commercial and business services to SOEs andthe private sector, the government could: a Reform budget management so as to better align budget resources with the priorities of government. While the central government level has made major progress in budget reforms in recent years, many local governments still follow a largely incremental process that does not allow retrenchment from areas that are no longer a priority. a Realign intergovernmental fiscal relations to create the preconditions for more equitable access to services. This would require a more explicit assignment of functions to each level of government and more equitable distribution o f fiscal resources among sub-national governments. X Deepening Public Service Unit Reform to ImproveService Delivery Reorganize the PSUsector 11. Determining the appropriate organizational form for providers of each service will require detailed review of the economic and technical nature of the service in question, as well as the environment in which service providers operate. However, China i s likely to benefit from: a Creating two categories of organizational forms for existing PSUs by making a legal distinction between devolved and delegated PSUs. Delegated PSUs could be defined as those that are directly affiliated with a core government department without an independent legal status and accounting system. Devolved PSUs on the other hand could be defined as those having an independent legal status and accounting system. a Establishing a legal and regulatory framework to allow for a broader variety of providers. In light of international experience, China could consider seven feasible organizational forms for service providers: (i)core government departments; (ii) delegated government PSUs; (iii) devolved government PSUs; (iv) SOEs in the public sector; (v) NPOs (min-ban PSUs); (vi) private small and medium firms; and (vii) largejoint stock companies outside the public sector. Strengthen accountability 12. In view of China's particular situation, PSU reform could aim at substantial strengthening in accountability through the following reformactions: a Establishing and enforcing pe~ormancemanagement compacts. Within the government, the central government could establish service standards for each ministry, sector, local government, and individual PSU. In addition to the realignment of intergovernmental fiscal relations, division of decision-making power among various levels of the government could also be reviewed to ensure that service delivery responsibilities are matched with adequate resources and decision-making power. Between PSUs and the government departments responsible for monitoring the performance of PSUs, resource agreements could be established to specify performance standards and rules on the use of budgetary and self-generated funds, including budgeting and treasury management. a Strengthening financial accountability. All inputs of public resources, most notably, land and fixed assets, subsidies, tax preferences and intangible assets, should be included in PSU cost accounting and reporting to produce a clear idea of exactly how much public resource has been invested in a particular PSU in a certain period. The management system for "administrative and PSU (xingzhen shiye Xing)" state assets needs a major reform to set up a functioning framework. User charges on key services such as education and health should be limited. a Deepening the reform of personnel and compensation system. Inlight of service standards and resource agreements mentioned above, each PSU should review its staffing level and consider options for retrenchment and reassignment, including early retirement, and voluntary separation. A realistic reform plan needs to be Executive Summan, xi designed to clean up all informal salaries, bonuses and other compensation by eliminating fraudulent and illegal incomes and regularizing others. 0 Empowering clients. Involving the public in service delivery monitoring and evaluation i s one method of empowerment. It is also worth considering the introduction of competition among providers by shifting financing from direct subsidization of providers (such as schools) to funding of the users (such as parents), and allowing them the choice of provider. Organize Reforms 13. The central authorities need to give considerable thought to the organization, principles, planning, and phasing of the reforms. Central guidance, coordination, and capacity for reforms are highly desirable. To organize the reform for better results, the following actions can be considered: 0 Setting up a taskforce. As a first step, the central government could set up a task force for PSU reforms directly under the State Council. The task force should, among other things: (i) carry out a survey of PSUs to deepen the information available on PSUs; (ii) design an overall strategy that includes principles for reform, phasing, timing, and responsibility for reforms; (iii) prepare draft laws, regulations and proposals for changes in legal rules; (iv) prepare the framework for performance management; and (v) coordinate the reform actions of ministries and local government. 0 Adopting a more decentralized review and classification approach. To better reflect the complexity and diversity of PSUs in a reform strategy, the government i s advised to adopt a more decentralized review and classification approach to replace the traditional classification approach adopted in past reforms, which does not differentiate the role of the government across services of varying economic nature, and i s not capable of capturing the sectoral andregional diversities. Manage the impact on employees 14. It is probably inevitable that some reform measures that are necessary for the improvement of public services will negatively affect the job, salary, bonus, and pension benefits of some PSU employees, at least in the short term. Managing the impact on employees i s therefore critical for the success of the reforms, including: 0 Communicating with employees on the limited scale of potential layoffs. Based on the nature of services provided by PSUs and the characteristics of its labor force, it i s conceivable that the government i s not facing the pressure of a mass layoff at the aggregate level or devastating employment pressure in certain regions, as was the case in SOE reforms. 0 Adopting a gradual approach to rationalizing the PSU labor force. One option could be to follow the pilot initiatives to explicitly introduce employment contracts, with potentially redundant employees being treated in some differentiated ways. xii Deepening Public Service UnitReform to Improve Service Delivery Being both determined and realistic in strengthening financial accountability. Strengthening financial accountability and deepening the reform of personnel and compensation systems may have a negative impact on some PSU employees. The PSU reform task force and the government face the challenge of designing a realistic approach that minimizes resistance to reform from potential losers, andin the meantime stays firmly committed to implementation o f the reform. A comprehensive review of pay scales in the public sector could provide a way forward. Aiming at a unified pension system. Inthe long run, a consistent national pension policy framework that applies to the entirety of the Chinese labor force inthe long term, including civil servants and PSU staff, i s the right direction of reform. Duringthe transition period, dual systems and parallel reforms will need to be in operation simultaneously. 1. CHINA'S PUBLIC SERVICEUNITSAND THEIR REFORM: AN INTRODUCTION WHATARE PSUsAND HowDoTHEYOPERATE? 1.1 The "Public Service Unit" (PSU), or shiye danwei, i s one of the four categories of public sector institutions or danwei in China.' Most shiye danwei are created to be public service providem2 A State Council decree in 1998 (State Council 1998a) defines a PSU as a "social service organization established by the state for the purpose of social public benefit;'' it i s created "by a state organ or other organization with state-owned assets" and carries out activities in education, science and technology, culture, health, or in other areas. 1.2 China has over one million PSUs (see Box 1.1)with a total employment of nearly 30 million in 2002 (NDRC PSU Team 20043), accounting for 41 percent of its public sector employment and 4 percent of total labor force. A substantial portion of China's economic resources i s devoted to PSUs, including 60 percent of its best-educated professionals, a large amount of state-owned land, around two-thirds of "non- commercial (fei jingying Xing)" state-owned assets, and one-third of the recurrent expenditures of the consolidated budget of all governments. 1.3 PSUs are highly concentrated in a few key sectors. Education and health together accounted for 70 percent of PSU employment in 2002 according to N B S data (NBS 2003a, Table 5-1, 5-7), or 53 percent according to NDRC PSU Team (2004). Scientific and technological research, cultural services and agriculture extension services are perhaps the next three largest sectors. The high degree of concentration in a few sectors, however, comes together with extreme diversity in terms of services provided. Consumers of their services include the general public, individuals, farmers, other business firms, government departments and the Communist Party. The other three are Communist Party or government departments (dangzheng jiguan), state-owned enterprises (SOEs) (including state-owned financial institutions), and state-sponsored social organizations (shetuan). Thus the translation of shiye danwei into "public service unit." Alternative translations include "public institutions," "institutional units" and "non-profit organizations." Unless specified otherwise, data used inthis chapter are from NDRC PSU Study Team, 2004. Data from Ministry of Personnel. The ministry has its definition of "professionals (zhuanye jishu renyuan)," which include most intellectuals such as teachers, doctors, scientists, engineers, actors, writers, etc. 1 2 Deepening Public Service UnitReform to Improve Service Delivery Box 1.1 LimitedAvailability of Dataon PSUs Despite the vast number and important role of PSUs in the economy, data on them are only limitedly available. The National Bureau of Statistics (NBS) and Ministry of Finance (MOF) are two key departments that collect data on PSUs. N B S has two sets o f data in the public domain. In its China Statistical Yearbook N B S releases annually the number of "staff and workers" of state-owned units that are explicitly defined as PSUs. Its Second National Census of Basic Units collected data on the number of units and employees of all "PSU legal persons" grouped by four-digit sub-sector and province as of the end of 2001. Starting from 1999, M O F has built up a database from the financial statements submitted by all PSUs which are independent accounting units5. This database contains unit-level information of key variables of revenue and expenditure of PSUs, including remuneration and social security expenditures. However, it has not been made publicly accessible. Other central government agencies also collect data on PSUs. Inparticular, the State Commission Office for Public Sector Reform (SCOPSR, or the State Office for Post and Establishment) collects data on the number and employment o f PSUs by sector and region. The Ministry o f Personnel (MOP) has a similar dataset, and neither agency has made its datasets publicly available. In addition, line ministries such as the Ministry o f Education (MOE) and the Ministry of Health (MOH) have their own databases focused on their sectors. Some of the data are published in their statistical yearbooks. This study has benefited from data of a report prepared by the PSU Study Team of the National Development and Reform Commission (NDRC) Source: NBS 2003a; NBS 2003b; NDRC PSU Team2004; World Bank. 2002a. 1.4 Government departments, PSUs and SOEs are often intertwined, making the boundary between PSUs and other public sector institutions blurred. In particular, it i s common for PSUs to have their own business entities, which may include listed companies controlled by universities and research institutes, as well as other less visible business operations. The boundary between the government and PSUs i s fluid as well. In past efforts to restructure government, staffs have often been moved to PSUs when pressures build to downsize government. In other instances, when government recruitment i s constrained by post and establishment controls, PSUs are sometimes created to circumvent restrictions. The relatively weak financial discipline of PSUs also created incentives for some government departments to use PSUs as vehicles for their revenue-generating activities. 1.5 PSUs are run by the government. They are affiliated with, and supervised by, authorities at any of the following six levels: (i) the State Council6; (ii)central ministry; a (iii)provincial(ormunicipal)government; (iv)aprefecture(ormunicipal)government; a (v) a county (or municipal, or district) government; and (vi) a township government. The affiliation and supervisory relationship i s determined when they are created. Organs of the Communist Party of China (CPC) are involved directly in supervision of some PSUs, such as those of the mass media. According to the State Council Decree 252 in 1998, when a PSU is created at a particular level of the government, one department of the government, the "approving authority," will approve the establishment of the PSU, which will then get itself registered with the Offices for Posts and Establishments of the See Handbookfor Compilation of Final Statement of Budgetary Unit, 2001,by Statistical and Evaluation Department, Ministry of Finance. Beijing: ChinaFinancial and Economic Press. In 2004, there were 14 PSUs directly reporting to the State Council, including, for example, China Academy o f Science, China Academy of Social Sciences, Development Research Center of the State Council, Xinhua News Agency. China's Public Service Units and Their Reform: An Introduction 3 appropriate level of government, and one department of the government will act as the PSU's "supervisory department (zhuguan bumen)." The latter often holds the power to appoint the management of the PSU, review and approve its budgetary, financial and staffing plans, and evaluate its performance. Most PSUs operate in rural areas of the country. Data from NDRC PSU Team (2004) indicate that nearly 65 percent of PSU employees are found in PSUs affiliated with county and township governments, which accounted for 83 percent of the total number of PSUs. 1.6 PSUs are typically partially funded by the budget. In 2002, budget allocations accounted for 47.6 percent of the total funding of PSUs. Another 48.2 percent of their funding was raised through charging fees, including "PSU service revenue (shiye shouru)" (45.5 percent) and "business incomes (jingying shouru)" (2.7 percent). Central and provincial PSUs tend to have more funding raised through business activities, while PSUs at the county and township levels are more reliant on budget allocations. Staff costs accounted for 29 percent of PSUs total expenditure in 2002, followed by other regular expenses (43 percent) and project investment (22 percent). The share of staff cost in total expenditure at the township level is three times that of the central level; at the county level it i s twice that of the central level. 1.7 PSUstaff share the same personnel system with civil servants inmost aspects: on recruitment, pension, and basic remuneration. Despite recent reform efforts, PSU staff enjoy defacto life-long employment. Their formal salary scales are set in line with those for civil servants. It is the greater flexibility of informal salaries and bonuses that differentiate PSU staff from their civil servant counterparts. PSU staff and civil servants also share the same pension scheme and medical insurance scheme, which differ from those for SOE employees. 1.8 Board governance i s uncommon in PSUs. The CPC Committee in each PSU i s often the most powerful decision-making body within the PSU, and the Party Secretary and the chief manager of a PSU are usually the people in charge. However, supervisory departments often retain significant influence over the decision-making processes of PSUs, although it i s not unusual for a PSU to find itself completely controlled by either its Party Secretary or the chief manager or both. PSUs follow special accounting and financial management rules set out by the MOF (MOF 1996), and are supposed to deliver their audited financial statements to MOF or bureaus of finance of the government with which they are affiliated. HOW HAVEPSUSBEENREFORMEDINTHE PAST? 1.9 A wide range of measures have been taken since the mid-1980s in reforming PSUs, and these represent an important part of China's overall reform program. In retrospect, the past reform efforts have some common characteristics. First, they have been made in the absence of a well-developed overall strategy. Second, the design of reforms has been often guided by classifications of PSUs based on the notion of "public benefit relevance." Third, the overall direction of reform has been market-oriented. Since the late-l990s, however, some new trends have emerged. On the one hand, private participation in public service delivery has increased; on the other hand, efforts at introducing market-like incentives into PSUs have been mixed, with the emphasis on the notion of "not-for-profit" in public service delivery. 4 Deepening Public Service Unit Reform to ImproveService Delivery Overall strategy and classification 1.10 It has been a consensus among policymakers and advisors that past PSU reform efforts have suffered greatly from the lack of a well-developed overall strategy and coordination. Some of them have characterized the past PSU reform as "touching the stones to cross the river without knowing where the other side of the river is" (Cheng Siwei 2001, p. 3), or "a reform without general objectives" (MOST 2003). These are very much fair assessments, as most reform measures have been implemented in PSUs by either line ministries (e.g., MOE, MOH) and provincial and municipal governments, or functional ministries (such as MOF and MOP) in more or less ad hoc and piecemeal manner. The overall reform strategy remains underdeveloped. The 10thFive-Year Plan, adopted by the National People's Congress (NPC) in March 2001, only required an acceleration o f reform of those social services that are "suitable to be run as industries (chanye hua)," as well as separation of PSUs from government and enterprises, and separation of for-profit organization from not-for-profit organizations. The recent Decision on Some Issues Concerning Perfecting the Socialist Market Economic Systems adopted by the CPC Central Committee in October 2003, a document that i s designed to provide overall guidelines for economic reform in the coming ten years, does state that the government will "implement PSU reform continuously," but specifies little more than reforms of personnel and compensation systems. 1.11 Since the very beginning of PSU reform, there has been strong emphasis on classification, reflecting the government's awareness of the high degree of diversity and complexity of PSUs, and the implication for the role of the government. As a result, reform measures have been differentiated across categories. The central concept of classification has been "public benefit relevance (gongyi Xing)." Despite the lack of clear definition of this term, it seems widely accepted that the government should take responsibility for services with high public-benefit relevance, while some others can be left to the market. It has been common to divide PSUs into three basic categories. On one extreme are units with law and regulation enforcement functions; on the other extreme are those whose services are considered of commercial and business nature. In between lies the main category of "public benefits-related units." Reform measures are typically differentiated on this basis. Pushinginto the market 1.12 Despite the vagueness of stated strategy and the ad hoc manner of implementation, the general direction of reform i s still discernible from actions taken over the past two decades. Overall, the reform has pushedPSUs further from government departments and closer to SOEs, a general direction of marketization, or as it has often been called, "pushing into the market." The central measure of reform along this direction is cutting budget funding to PSUs to force them generate their own revenues to survive. "Contract responsibility systems" of various forms were introduced to implement "budget contracting (yusuan baogan)," whereby the government allocated a predetermined amount of funds to PSUs, which were then supposed to retain any surpluses or make up any funding gaps. To facilitate revenue generation activities, managers and team leaders inresearch institutes, universities, schools, hospitals andother PSUs were granted a great deal of autonomy with regard to operational decision-making such as planning, budgeting, China's Public Service Unitsand Their Reform: An Introduction 5 spending of self-generated revenues in investment and incentive schemes for employees, creating and abolishing internal departments, setting prices for services that are not regulated, borrowing from banks, etc. However, appointment of managers remains a power held by the supervisory department and related CPC organs, and few PSUs have been granted the full set of decision-making rights inhiringandfiring. 1.13 In attempting to introduce market-like incentives, the reform has restructured the remuneration system in some complicated ways over the past two decades. The general direction has been one of "growing out of' the old system. Instead of scrapping the old system of fixed salaries based on qualifications, the reform added to it a new "flexible" component, in the hopes that performance can be better reflected in compensation. Government control over the "fixed" component remains tight, at least nominally, while the "flexible" component, which i s often funded with self-generated revenue, has been largely influenced by the labor market. In the meantime, more flexibility was also granted to professionals to realize their full potential by providing services to clients for fees without having to quit their jobs inPSUs. 1.14 When a PSU i s completely "pushed into the market," it operates in almost the same way as a Chinese SOE. This gave rise to a formal category of PSUs that are "managed as enterprises (shiye danwei qiyehua guanli)." In recent years, a significant number of PSUs, including some "PSUs managed as enterprises," have been formally transformed into enterprises, either wholly state-owned or under mixed ownership. Some have been privatized. This i s particularly the case for those that have been classified as "commercial and business PSUs." Increasedprivateparticipation 1.15 Inaddition to the marketization of PSUs, increased private participation inservice delivery has become an important part of PSU reform efforts in recent years, as it exposed PSUs to some degree of competition with non-PSU providers. The non-state sector counterparts of PSUs are named as "min-ban non-enterprise units (MNEUs)" (State Council 1998b), defined as "social organizations that are created by enterprises, PSUs, social associations, other entities and individual citizens with non-state owned assets, and engaged in not-for-profit social services activities.'' According to N B S data, there were 0.12 million MNEUs in 2001 with 1.2 million employees, equivalent to 4.5 percent of total PSU employment (NBS 2003b). 1.16 In the education sector, the expansion of "min-ban (private)" institutions since 1996 has been spectacular, with total enrollment increasing 2.3 times, but their share of total enrollment only managed to reach about three percent in 2002. However, this does not include min-ban kindergartens, which enjoyed a much bigger market share, nearly five percent in 1996 and 20 percent in 2002 (Wang Rong 2004, p. 7). Min-ban education institutions have developed in market niches left uncovered by the public sector. This includes not only "elite" schools, but also min-ban schools created in response to the demand of the poor and disadvantaged, where the public sector fails to live up to its promises. In the health sector, private participation remains low as well and, unlike the education sector, shows no expansion trend. Interms of beds and employees, the 1990s saw little increase in the share of private providers (Li Ling, 2004, pp. 13, 17). The Chinese health sector remains overwhelmingly state-owned. 6 DeepeningPublic Service UnitReform to Improve Service Delivery 1.17 Privatization of education and health PSUs has been widespread in recent years. While publicly-available data does not allow any systematic assessment, anecdotal evidence does exist to suggest the nature and methods of privatization, known inChina as "transformation (zhuan zhi)." In the education sector, privatization of public schools and universities has taken some complicated approaches. "Transformed schools (zhuan zhi xue xiao)" are not fully privatized schools in typical cases. They are "publicly-owned and privately-run," i.e., their assets remain state-owned and their staff are still PSU employees. However, the management has been granted such extensive autonomy that these schools are managed like private schools interms of fee collection, hiring andfiring staff, recruitment of students, management of assets and logistic services, etc. For higher education institutions (HEIs), the vehicle of privatization i s "independent colleges," which are subsidiaries of existing public H E I s with "new mechanisms and new models," as specified by an MOE circular in 2003.7 The circular allows enterprises, PSUs, social organizations and private individuals to run independent colleges as partners of their parent public HEIs. It requires all independent colleges to adopt "min-ban mechanisms" and a "new model of operation." 1.18 Privatization initiatives in the health sector bear a clear resemblance to SOE reform. Private ownership has been brought in through a variety of ways. Management and employee buyouts (MEBOs) are one of them. Sales to outsiders have also taken place. Joint venture i s another way whereby a public hospital can be restructured by foreign investors. Box 1.2 provides information about one of the most controversial cases of hospital privatization, but indeed many provinces and cities, including Beijing and Shanghai, have announced their plans to corporatize public hospitals and bring in private ownership. In some places, hospitals that were detached from SOEs were privatized first, as they are often in financial difficulties and in need of additional investment for restructuring. Introduction of the notion of "not-for-profit" 1.19 Since the late-l990s, non-profit organizations (NPOs) have been seen by many policymakers and advisors as an alternative to reforming incentives of PSUs. The NPO as an organizational form i s not defined in Chinese law. Nonetheless, the notion of "not- for-profit" has made its way into policy-making and legislation. This largely reflects the efforts to shape incentives of service providers by introducing the notion of not-for-profit. The sectors of education, health and science and technology (S&T) research are the three most visible examples. China's Public Service Units and Their Reform: An Introduction 7 I Box 1.2 Hospital Privatization inSuqian City, Jiangsu province I Suqian city of Jiangsu province has been in the center of a debate for its privatization o f all hospitals. In this rich province, Suqian is a relatively poor city with a population of 5.15 million and GDP of RMB4,500 per capita in 2001. Starting in 2000, the city privatized all 124 township level and 6 county level public hospitals. The total amount of the sales was RMB323 million. Newspaper reports described the privatization as having led to declined service standards and higher prices. During 2000-02, average outpatient fees rose from RMB90 to RMB113 in city level hospitals, from RMB51to RMB143 in county level hospitals, and from RMB32 to RMB35 in township level hospitals. Some private owners found themselves in financial difficulties because the hospital assets were over-valued or liabilities were hidden. In one county hospital bought by a doctor with RMB130 million, even the ambulance department was asked to generate its own revenue to cover costs. ISource: China Business Post (CaQingShibao),September 20-26, 2003. I 1.20 Inthe education sector, the Education Law of 1995 requires that "no organization or individual shall run schools or other educational institutions with a purpose of profit- making" (article 25). The Private Education Promotion Law, passed in 2002 after intensive debate, states that "after deducting running costs of the school, setting aside a development fund and other provisions according to relevant government regulations, the investors of the private school may extract a reasonable return out of the surplus. Specific rules on extraction of a reasonable return shall be formulated by the State Council." Such specific rules were enacted on April 1, 2004 by the State Council in its Regulation on Enforcement of the Private Education Promotion Law.8 An important step taken by this regulation i s that it allows each min-ban institution to declare if it chooses to claim or not claim a "reasonable return." 1.21 While the Education Law only mentions "purpose of making profit," the concept of NPO as an institution was explicitly introduced into China's health sector in 2000 in a reform program designed jointly by eight ministries led by former the State Council Office for Restructuring Economic Systems (SCORES) and approved by the State Council (State Council 2000). The program calls for the establishment of a "new management system of medical and health institutions" in which all institutions in the sector are classified as either NPO or for-profit organizations, with the former being expected to play a "leading" role. 1.22 The S&T sector i s another sector where the NPO as an institution has been introduced into a reform program inan explicit manner. The CPC Central Committee and State Council Decision on Strengthening Technological Innovation, Developing High- Tech and Promoting Industrialization (Document No. 14)9in 1999 adopted the NPO as an institutional arrangement for reform of some S&T research institutions. It stated that "those that provide public services to the society and cannot expect adequate economic return should be run and managed as NPOs." On April 29, 2000, the State Council approved an implementation plan formulated by 12 ministries led by MOST (Notice No. 38). The plan further specifies that the model of NPO applies to those institutes that http://www.moe.gov.cdj yfg/laws/mbcj.htm http://www.most.gov.cdShowContent.jsp?db=zczdjd~35&id=6 8 Deepening Public Service Unit Reform to Improve Service Delivery "engage mainly in application-oriented basic research or public service delivery, cannot expect adequate economic return, really need government support." WHAT ARE THE PROBLEMSNOW? 1.23 PSU performance i s crucial for improving service delivery in the public sector to a level commensurate with a "xiaokang " society and "people-centered" development. Past reform efforts have achieved much, but at the same time introduced undesirable incentives in services delivery. As a result, service quality is highly uneven, cost- effectiveness i s falling short, and access to services i s inequitable. Key services are underfunded; vulnerable social groups with lower income are burdened with rampant and excessive user charges, and yet, financing PSUs remains a heavy burden to government at every level. PSUs have strong incentives for "revenue generation" but much weaker ones to provide high quality services at low cost or, even worse, incentives for revenue generation have been causing deteriorating quality and producing inequitable outcomes in some areas. Moreover, the large and omnipresent PSU sector i s hampering the emergence of a non-state service sector crucial for China's continued high growth and productivity. Improving service delivery will require a far-reaching reform, which will have to address both financing andincentive problems in a coordinated manner. Financing 1.24 The challenge of financing services for China comprises two parts. The first involves a reshaping of the public-private divide of financing responsibility: the government must meet the needs of financing services for a well-off society while ensuring that the financing costs do not exceed the capacity o f its resources. The challenge for the reform i s to make those tough choices in adjusting the range of services offered to fit within the level of available public resources, but doing it in a way that protects sufficient financing for the services in which the government's role i s critical in correcting market failures, as well as achieving equity goals. 1.25 Viewed from this perspective, the past two decades have witnessed a privatization of service financing: a shift of financing responsibility from the government to private users. This i s especially clear in the cases of education and health, as shown by Figure 1.1 and Figure 1.2. In the education sector, this is reflected by the sharp increase in tuition fees in relation to total revenue of education institutions. In the health sector, it resulted in the rising trend of out-of-pocket expenses as part of total health expenditure, given the relatively underdeveloped insurance system. 1.26 In recent years, however, the government has made very large additional injections of funds into PSUs. From 2000 to 2002, budget appropriations to PSUs were raised from RMB362 billion to 659 billion, an increase of 82 percent." (This raised the share of budgetary spending on PSUs from 4.1 to 6.4 percent o f GDP.) Yet the increased funding did not appear to have produced a corresponding decline in fees except in a few services where fee reductions were mandated by government (e.g., in compulsory lo thiswasaperiodofmilddeflation, theseincreasesareapproximatelyinrealterms. Since China's Public Service Units and Their Reform: An Introduction 9 education). During the same period self-raised revenues of PSUs grew 43 percent and "service revenues" grew by 53 percent. Figure 1.1 The Evolutionof Education Financing in China, 1986-2002 1988 1887 1988 1989 1990 1981 1992 1993 1994 1995 1996 1997 1998 1998 2000 2001 2002 Year /-Total educational revenue +Government funding +Tuitions and fees1 Source: Finance Department of MOE and Social Department of NBS, China Educational Finance Statistical Yearbook, variousyears. Beijing: China Statistics Press. MOE Finance Department and ShanghaiIntelligence Institute, 1997. "Annual Report of National Education Finance Development." Citedfrom Wang Rong 2004. Figure 1.2 The Evolutionof Health ExpendituresinChina, 1978-2001 1.27 Many signs point to the additional resources being increasingly captured by PSUs for increasing employment and wages. Table 1.1 shows personnel expenditures (including basic wages, subsidies, bonuses and social security contributions, as well as pension payments) growing faster than budgetary appropriations for PSUs over the 2000- 2002 period. With increased employment, total personnel expenditures grew by 91 percent, outstripping the growth inbudget appropriations. This phenomenon i s especially marked at the lower levels of government, where total personnel expenditures outgrew 10 Deepening Public Service Unit Reform to Improve Service Delively budget appropriations by substantial margins and are likely to be squeezing out other expenditures. Table 1.1 Growth of the PSUPersonnel Expenditure (%), 2000-200211 Number of Average Personnel Budget Employees remuneration Expenditures Appropriations National Totals 23 25 91 82 Central -6 25 55 71 Provincial 10 1 86 91 Prefecture 7 28 84 35 County 84 27 107 85 Township 1 12 80 61 1.28 This leads to the second part of the challenge: to find a smarter strategy that enables the government to achieve more of its social objectives out of a given amount of public spending, or to get more "bang for the buck." In this regard, two problems are outstanding. 1.29 Weak budgeting institutions. In allocating limited public resources, all levels of the government need to establish and implement their orders of priority in financing services to ensure the realization of their policy objectives. In this sense, the existing budgeting institutions are weak. Budgetary spending does not appear to be aligned to policy priorities of the government as stated in government documents. There does not appear to be any discernible public expenditure prioritization, either on efficiency grounds or on equity grounds. Indeed, not every part of public spending i s accounted for. In addition to budgetary outlays, the government provides non-budgetary support to PSUs through several other channels: the free use of state-owned land and physical capital, the use of invisible capital granted by the government, and tax expenditures in various forms of preferential "policies." This non-budgetary part of public spending on PSUs appears to be non-transparent not only to the public, but also to the government itself. As a result, the financial operations of PSUs are substantially beyond the monitoring and control of the government. The informational basis for sound budgeting i s lacking. 1.30 The inability to hold PSUs accountable for their spending of public funds i s more evidence of weak budgeting institutions, and these are reflective of wider problems associated with the budget system in China. In spite of significant efforts over the past ten years, budget reform has lagged behind other reforms in the Chinese economy. As the economy became increasingly decentralized and market-oriented, the increasingly inappropriate budgeting process became a major constraint to the effective use of the budget for resource allocation and macroeconomic management.l2Through the 199Os, the budgeting process was passive and incremental, had little focus on strategic priorities, l1 2002personnelcosts arecalculatedonanewbasis andareperhapsnot strictly comparable to The earlier figures. However, the rates of increase for 2000-01 and 2001-02 are quite similar in spite of this change. l2SeeWorld Bank(2000a) for anin-depth analysis ofChina's publicexpenditure management systemin the late-1990s. The report's analysis remains relevant today, since many o f the problems persist. China's Public Service Unitsand Their Reform: An Introduction 11 was unable to shift expenditures to emerging needs, and was unable to adequately track how public funds were spent, or even how many people the government employed. Budgeting for the PSUs came from this system, and it producedthe same inefficiencies. 1.3 1 Znadequate inter-government fiscal system. In addition to reforming budgeting institutions to optimize the delivery of services, the government has to institute an effective division of labor at every level of the government, and assign appropriate instruments of financing to match-whether through revenue assignments or transfers. By a variety of conventional measures, China i s the most decentralized country in the world. Sub-national governments account for 70 percent of total budgetary expenditures, with the bulk of that at the sub-provincial levels. Counties and townships together account for 55-60 percent of public spending on health, and 70 percent of that for education (World Bank 2002a). This highly decentralized system could be a boon to managing service delivery, but under the current arrangements, sub-national governments have inadequate revenues for meeting their heavy expenditure responsibilities, and receive little help from the system of intergovernmental transfers. As a result, they are to a large extent self- financed, as suggested by Figure 1.3 and Figure 1.4, and no mechanisms exist to ensure minimumservice standards across regions. As regional disparities in income grew, this led to growing regional disparities in services through the 199Os, and a default in the delivery of vital services in many poor localities, including education and health.13 1.32 This inadequate intergovernmental system has contributed to dilutingthe tripartite budgeting division among the PSUs (i.e., fully-funded, partially-funded and self- financed). Since the implementation of national priorities depends on local funding, even key services are often underfunded in poor r e g i ~ n s . 'This situation i s common to many ~ core public service providers such as primary schools, centers for disease control, and the like, all of which are financed by local budgets. When local budgets are in difficulty, these "fully-funded" PSUs are forced to develop fee-raising services like all others.15 A second aspect of this decentralized fiscal management, incombination with the budgeting practice that continues to allocate resources by employees, rather than programs, has encouraged and sustained the rapid employment growth in PSUs that i s hindering recent reform efforts to refocus public spending on priorities. The implicit understanding that higher-level governments will bail out lower-level governments on wage payment arrears distorts allocative decisions at local levels. The traditional control via the post and establishment offices appears increasingly ineffective, as the several rounds of wage increases for civil servants have raised their wages to very high levels relative to local income. l3 around1998,anumberofnewpolicieshavebeenintroducedtoalleviatefiscalproblemsatthe Since lowest tiers of government, but these have been piecemeal and have had limited effect. l4ThisisthekeyfeatureoftheChinesefiscal systemthat ledtheWorldBankreport, China:National Development and Subnational Finance (2002a), to state that the intergovernmental system constitutes a bottleneck to national policy implementation. l5 inefficientduplicationthatresultsfrommaternalchildhealthcentersbranchingouttodeliverbabies The and providing pediatric care, epidemic prevention centers offering curative care, etc., are discussed in World Bank 2002b. 12 Deepening Public Service Unit Reform to Improve Service Delivery Figure 1.3 EducationExpenditure Per Student and GDPPer Capita by Province, All Primary Schools, 2002 BOW - sow . I 0 5000 ,OW0 l5000 200w 2SWO 3WOO 35000 40000 45000 GDP per capita Source: China Educational FinanceStatistical Yearbook,2003. Figure 1.4 Education Expenditure Per Student and GDPPer Capita by Province, Rural Primary Schools, 2002 e 5000 cm a 4500 4000 0 5000 ,0001 lSW0 20WO 25000 3WOl 35000 4WW 45000 GDP per capita Source: ChinaEducational FinanceStatistical Yearbook,2003. Incentives 1.33 The way services are financed has its impact on the incentives of PSUs and their managers and employees. Overall, PSUs are characterized by strong motives of revenue generation and weak disciplines from either the government or clients. 1.34 Strong motives of revenue generation. The incentives for revenue-generation were created when the past reform "pushed PSUs into the market." As the reform cut budget funding to PSUs and gave them greater managerial autonomy to extend the scope of services and permitted them to link managerial and staff remuneration to the revenues raised, managers and employees of PSUs found themselves in the position o f residual claimants. These concessions were made at a time when civil service salaries were low and there was widespread acceptance of the need to raise them. Over time, however, "topping up salaries" has become institutionalized and built into the remuneration structure, creating the incentive for PSUs to seek "own revenues," and "own revenues" China's Public Service Units and Their Reform: An Introduction 13 came to play a role akin to profits for private firms operating in a market environment. Firstly, the "flexible" component of remuneration of managers and employees, in the forms of bonuses and subsidies, has grown so large as to account for about a half of the total remuneration on average, and they are most likely financed by "own revenue," as suggested by Table 1.2. Secondly, there i s probably a fairly strong correlation between the level of the "flexible" component of remuneration of employees of a PSUand its own revenue. Data show that such a strong correlation does exist at the provincial level. As shown inFigure 1.5, there i s wide variation across provinces in the levels of bonuses and subsidies, with the ratio of highest to lowest being 18.3 to 1. Regression analysis on provincial-level data found that the amount of own revenues per employee was an excellent predictor of average pay (with an R-square of 0.84), but no statistically significant relationship to the share of revenues derived from budget appropriations. Table 1.2 Bonus and Subsidies as Components of Remunerationof PSUEmployees, 2002 Averagebasic Average bonuses and Basic wageltotal Funding from wage subsidies remuneration (%) budgevtotal PSU funding (%) National average 6,238 4,807 56.5 47.6 By tier: Central 6,151 8,850 41.0 48.6 Provincial 6,320 6,879 47.9 44.4 Prefectural 6,95 1 6,653 51.1 38.1 County 6,030 3,478 63.4 53.3 Township 6,430 3,727 63.3 60.1 Figure 1.5 Fitted plot of average pay on "own revenues," by province, 2002 60,000 1 * io*ooo i --4 0 1 0 20,000 40,000 60,000 80,000 100,000 120,000 140,OO Own revenuedstaff 0 Actual Y - Redicted Y 1.35 To mobilize employees to generate revenue, it i s common for PSU managers to link the pay level of employees with some observable and verifiable indicators such as the amount of financial revenue they generate. Box 1.3 presents one example of a "performance-based" remuneration system as an illustration. 14 Deepening Public Service Unit Reform to Improve Service Delivery I Box 1.3 Remunerationof DoctorsinanUrbanHospital A local newspaper reported that a Chinese traditional medicine hospital implemented a new performance-focused compensation system in2003. Under this new system, doctors receive on average 30 percent of their regular salary as base pay. The remainder i s made flexible and dependent on "performance." A doctor in a "poorly performing" department receives RMB700per month while those in "better performing" departments may earn as much as RMB10,000, The best performing department turns out to be "Department of Occupational Injury Verification." Doctors also get a part of their pay based on the value of drugs and services they prescribe. For example, a doctor receives 11percent of the value of Chinese herbal medicines, 2 percent for western medicines, 10 percent for chemical and physical tests, 13 percent for injections, RMB17 for each CT test, andRMB10 for each Ultra Sound test. Source: Local newspaper report. See: http://news.sina.com.cdc/2004-06-07/0I 593389284.shtml 1.36 Weak disciplines. Strong revenue-generating motives bring about improvement in service delivery results only when they are met with equally strong market and regulatory disciplines. This i s particularly critical for services such as education, health care and agriculture extension services, which are mostly characterized by low "measurability," i.e., a high degree of difficulty in measuring the quantity and quality of the service with reasonable precision. To China's PSUs, disciplines from both the government and the market appear to be weak in relation to the strong motives of revenue generation. Past reform has indeed "pushed" most PSUs into the market, but the market itself i s a highly distorted one. Discipline from the market i s weak due to three primary institutional weaknesses: consumers are not sufficiently empowered; the regulatory framework remains underdeveloped; and market competition i s incomplete and often distorted. 1.37 The health sector i s an example. While high-powered incentives have been created to drive hospital managers and doctors to generate financial revenue, the market for their services i s such that this i s more difficult to achieve through improved quality of services, because on the one hand fees of services are tightly regulated and underpriced, and on the other hand, quality of services are often not measurable, verifiable and not necessarily appreciated by patients. As a result, the direction along which powerful incentives work i s to encourage hospitals to sell more drugs at high prices, which i s profitable because the market i s hardly competitive, and to skimp on quality of services wherever feasible (inparticular, those non-observable and non-measurable dimensions of quality) to save cost. Despite increasing tension between patients and health workers, there are good reasons to believe that the existing problems in the health sector could have been much more serious without the counter-functioning of the professional integrity and public service ethics demonstrated by the majority of health care professionals. 1.38 Discipline from the government has been also weak. This is largely a result of weak performance management. Under the existing system, it i s common for one government department to act as both policymaker, i.e., the government agency that i s supposed to monitor and evaluate the performance of service providers, and as service provider when it acts as the "supervisory department" of PSUs. This i s the problem known as "non-separation of supervision and management (guanban bu fen)." The risk of the policymaker function being captured by the service-provider function i s high and realistic. There i s a general inadequacy of institutionalized efforts in monitoring and evaluating service delivery results and the degree of satisfaction of citizens/consumers. Under such a system, it i s natural for service providers such as schools, hospitals, and China's Public Service Units and Their Reform: An Introduction 15 agriculture service stations to mind much more about the satisfaction of their "supervisory departments" than that of parents, patients andfarmers. WHAT REFORMSTRATEGIESHAVEBEENPROPOSED? 1.39 A few systematic studies have been conducted by academics and policy advisors in collaboration with relevant government departments in recent years. They have put forward proposals for future reform, drawing on China's experiences to date, as well as on theory and on international experience. Three major studies have been completed by the following scholars and agencies: (i)a research team led by Cheng Siwei; (ii) Development Research Center (DRC) of the State Council, in collaboration with SCOPSR and MOF; and (iii) a research center of MOST, again in collaboration with the MOF. In addition, NDRC i s leading an on-going multi-ministerial study on PSU reform strategy. Some of its key views were presented in an international seminar in March 2004. The ChengSiwei Study 1.40 The highly influential study prepared under the direction of Cheng Siwei16 emphasizes the need for a "target model" for PSU reform and goes on to propose a model based on six elements, the first of which i s the need to distinguish between public and private goods. The study then proposes to transform all those PSUs that do not produce public goods into enterprises, and those that are involved in public goods delivery into NPOs, which can be run either by the government or by the private sector. The government should take the main responsibility for running pure public benefit NPOs, while quasi- public benefit ones should mainly be run by the private sector. The second element of Cheng Siwei's "target model" i s the need to "integrate social resources," financing the provision of public goods through charitable donations and user charging. The study calls for "non-profit organizations becoming the main source of innovation in social services" in China, and for "revenues of non-profit organizations coming mainly from society." The remaining four elements o f Cheng' s "target model" relate to reforms he suggests are necessary to promote the successful emergence of non-profit organizations. These include the introduction of "preferential policies for non-profit organizations," including tax exemption for qualifying charitable institutions (element three); development of a strategy for "recruiting volunteers," on the grounds that Chinese non-profit organizations should rely more on volunteer staff (element four); enactment of "reinforcing legislation," including legislation to clarify the independence from government of non-profit organizations (element five); and implementation of "scientific management,'' including skills training for managers and staff of non-profit organizations (element six). Overall, this amounts to a proposal with NPOs at its center. The DRC study 1.41 The full report of the DRC study has not been published. However, a summary paper presented by ,Ge Yanfeng (2003), the team leader, raises some critical issues and illustrates some important differences o f emphasis from the Cheng Siwei study. In l6 EnglishlanguagesummaryofthereportformsChapter10ofChengSiwei(2001). An 16 Deepening Public Service Unit Reform to Improve Service Delivery particular, the DRC study argues that some PSUs providing "social services" with major "public benefits" have become "over market-oriented," with "excessive autonomy and lack of restraint." Profit maximization has become the sole objective, with such PSUs "ignoring social objectives or even harming social objectives in order to acquire economic gain." The study stresses that government must take responsibility for organizing such "social services," either directly through public ownership or indirectly. The study acknowledges a role for delivery through non-profit organizations, and for contracting with private for-profit participants. However, it warns of the danger of "putting the cart before the horse." Furthermore, the study warns, reformers should be suitably cautious about institutional innovations such as delivery through NPOs and make sure that increased autonomy of PSUs i s matched with accountability. Regardless of the mode of delivery, the study emphasizes, "government should be the sponsor" for "social services with public benefits," including providing the necessary budgetary inputs. This will involve re-prioritization of financing towards social services with public benefits. It will also involve a comprehensive reform of inter-governmental fiscal transfers, in order to ensure that assignment of revenues and transfers are adequate, given functional assignments to different tiers of government. The MOSTStudy 1.42 Like the Cheng Siwei team, the MOST study (MOST 2003) calls for the establishment of some "general objectives" for PSU reform, characterizing the reform experience to date as "groping along the cave wall." In other regards, however, the conclusions of the MOST team are closer to those of the DRC study than the Cheng study. 1.43 The MOST study suggests that commercial entities should become major players in the "private goods" part of the service sector. The study argues that both theoretical analysis and proven experience demonstrate that market-based approaches will often achieve greatest efficiency in the provision of private goods. However, the study goes on to argue that government must bear the major responsibility for ensuring the provision of public goods. It notes that government may either provide such public goods directly or else procure them, through budgetary appropriations, from non-governmental organizations (either the for-profit private sector or NPOs). The study suggests that NPOs have a "supplementary role" to play in the service sector. The study also includes a strong warning that "public service erosion" i s resulting from China's current strategy for PSU reform, which the authors characterize as "loosening the tight control" and "encouraging to make more earnings": "Some public institutions (PSUs) have ignored their mandated public goods services, with increased focus on making more money through business activities, causing the problem of the market eroding public services.'' The NDRCStudy 1.44 Views of the NDRC were outlined by Director General Fan Hengshan of the Comprehensive Reform Department of NDRC in his keynote speech (Fan Hengshan 2004) in an international seminar jointly sponsored by the NDRC, OECD and the World Bank on PSUreform.17 In sum, Fan proposed to redefine PSUs as organizations that fall l7 waslaterpublishedinChinaEconomicTimes,April12,2004. This China's Public Service Units and Their Reform: An Introduction 17 in between government and enterprises and engage in social services and public-benefit services. "Their basic features are non-government (nor secondary government), non- enterprise (nor quasi enterprise) andnon-profit (nor for-profit indisguise),'' as he put it. 1.45 Fan suggested that PSUs should operate in areas of "public causes, public benefit services, social mutual assistance, charitable assistance," etc. Accordingly, there can be three types of PSUs with varying functions: (i) PSUs engaging in social services and public utilities services, whose function i s to meet ever increasing social demand by promoting provision of public goods and services; (ii) PSUs engaging in social security and relief, as well as activities related to charitable courses, etc., whose function i s defined as promoting the development of social relief and the value of common prosperity; and (iii) PSUs serving as communication channels of the government and enterprises. They serve both government and enterprises by promoting the coordination of behaviors of the two parties. On financing, the following models were proposed: (a) Fully funded by the government. This applies to PSUs that provide needed pure public benefit social services. Examples include basic education, public health, basic science and technological research. Generally, this kind of PSU should be runby the government. (b) Donations. PSUs engaging in social relief may set aside a proportion of the donations they receive to finance their operation, subject to strict auditing and monitoring. (c) Government contracts. Government may contract out some of the services it wishes to deliver. (d) User charge. Those serving as intermediaries between the government and enterprises may acquire revenue by charging a fee for their services. 2. INTERNATIONALEXPERIENCEINPUBLICSERVICE PROVISIONAND ITSREFORM THEFINANCEOF PUBLIC SERVICES IN OECDCOUNTRIES 2.1 Interms of financing public services, the experience of most OECD countries points to a growing role of the state vis-&-vis the private market. As Table 2.1 illustrates, General Government Expenditure as a percentage of GDP increased significantly in all OECD countries over the 20th century-from an average of 12.1 percent of GDP to 45.4 percent of GDP between 1913 and 1996 in the 13 sampled countries-although it has tended to stabilize since the 1980s. Despite the dramatic changes in ownership structure over the past two decades, it i s notable in Table 2.2 that the share of GDP allocated as General Government Expenditure to different sectors has changed rather little in OECD countries over this period. Indeed, General Government Expenditure on public education has increased from an average across 17 western countries of 5.8 percent of GDP in 1980 to 6.1 percent in 1996; and on healthcare from 5.8 percent of GDP in 1980 to 6.4 percent in 1996. A more recent data source suggests that the public share of health expenditure remained high and virtually unchanged in most countries during recent years-accounting for 72.3 percent of total health spending on average across OECD countries in 1997 and 73.1 percent in 2002." Likewise, the public share of financing for schools, universities and other educational institutions remains high in OECD countries: over 88 percent of all funds for educational institutions, on average, and 93 percent for primary, secondary and post-secondary non-tertiary levels of education (Pilichowski, 2004). 2.2 What the data show i s a fundamental transformation in the role of the state in western countries over the 20th century. The state was once not much more than a "watchman" with such responsibilities as protecting against outside threats, maintaining law and order, and carrying out housekeeping services; nowadays, the state i s defined in terms of the financial security it provides individuals and the life opportunities it provides them through enhanced education and health services. 2.3 Asian OECD countries such as Korea and Japan maintained lower shares of expenditure in GDP than western OECD countries. However, they also enjoyed a more equitable pattern of economic development compared with the western OECD countries (and compared with China inrecent years). This may have reduced the need to subsidize access to public services and to provide transfers for equity reasons. Furthermore, spending has increased as a share of GDP as these economies have matured. '* OECD (2004). Total health expenditure as a percentage of GDP increased on average across OECD countries over the same period, from an average 7.8% in 1997 to 8.5% in2002. 19 20 Deepening Public Service Unit Reform to Improve Service Delivery Table 2.1: GeneralGovernmentExpenditure as Percentage of GDP, 13 OECD Countries, SelectedYears, 1913-1996 Country 1913 1920 1937 1960 1980 1990 1996 Australia 16.5 19.3 14.8 21.2 34.1 34.9 35.9 Canada d a 16.7 25.0 28.6 38.8 46.0 44.7 France 17.0 27.6 29.0 34.6 46.1 49.8 55.0 Germany 14.8 25.0 34.1 32.4 47.9 45.1 49.1 Italy 17.1 30.1 31.1 30.1 42.1 53.4 52.7 Japan 8.3 14.8 25.4 17.5 32.0 31.3 35.9 Netherlands* 9.0 13.5 19.0 33.7 55.8 54.1 49.3 New Zealand d a 24.6 25.3 26.9 38.1 41.3 34.7 Norway 9.3 16.0 11.8 29.9 43.8 54.9 49.2 Spain * 11.0 8.3 13.2 18.8 32.2 42.0 43.7 Sweden 10.4 10.9 16.5 31.0 60.1 59.1 64.2 United Kingdom 12.7 26.2 30.0 32.2 43.0 39.9 43.0 United States 7.5 12.1 19.7 27.0 31.4 32.8 32.4 Average 12.1 18.9 22.7 28.0 42.0 45.0 45.4 Source: Vito Tanzi and Ludger Schuknecht,Public SDendina in the 2OthCenturv, Cambridge University Press, 2000. Notes: * Through 1937,data cover only central govemment expenditures. Table 2.2: General Expenditure as Percentage of GDP, by Sector, 17 Western Countries** Averages, SelectedYears, 1870-1996" Expenditure 1870 1913 1937 1960 1980 1996 ConsumptionExpenditure 4.6 d a 11.4 12.6 17.9 17.3 Subsidies and Transfers 1.1 d a 4.5 9.7 21.4 23.2 Public Education 0.6 1.3 2.1 3.5 5.8 6.1 Health d a 0.3 0.4 2.4 5.8 6.4 Pensions d a 0.4 1.9 4.5 8.4 9.6 UnemploymentPayments d a d a 1.3 0.3 0.9 1.6 Public Investment 2.0 2.8 3.8 3.2 3.5 2.9 Defense d a d a 3.7 3.4 2.5 2.0 Source: Vito Tanzi and Ludger Schuknecht,various tables. Notes: * Year indicated, or closestyear for which data are available. ** The I7 countries are: Australia, Austria, Belgium, Canada, France, Germany, Italy, Ireland, Japan, Netherlands, New Zealand, Norway, Spain, Sweden, Switzerland, United Kingdom, and United States. PUBLICSERVICE PROVIDERSOECDCOUNTRIES IN 2.4 While shiye dunwei i s a uniquely Chinese institution, there are public service providers inevery country, referredto and organized indifferent ways but playing similar roles in their economies. Compared with PSUs, public service providers in OECD countries show a wide variety of organizational form, budgetary relations, funding, and International Experience in Public Service Provision and its Reform 21 governance. Inparticular, there i s considerable involvement of non-state organizations in the production of public services, even if financed from the budget. And in sharp contrast to PSUs in China, there are few public agencies that are involved in predominantly commercial activities. Within the public sector, the legal form that a public service provider can take i s more diverse than the all-purpose model of PSUs in China. 2.5 These general points notwithstanding, it i s important to realize that there i s not a single "OECD model," and the full picture of public service providers in OECD countries i s characterized by high degree of diversity and complexity, which are partly reflected in the different ways they are namedin different countries (Table 2.3). "Overall, it is poorly defined territory," as an OECD study has found (OECD 2002, p.11). There i s no universally accepted classification of these entities. Table 2.4 provides a list of some discernible types of organizational forms of these entities, as well as providers in the private sector and the NPO sector. Their particular features are further described in some detail inBox 2.1. Table 2.3 Namesof Agencies, Authorities and Other Government Bodies Involvedin Service Delivery inSome OECD Member Countries Country Names ~ Canada Service agencies, Special Operating Agencies (SOAs), departmental service agencies, shared governance corporations France Public establishments, independent administrative authorities Germany Federal agencies (direct federal administration), bodies of public law (indirect federal administration) Netherlands Independent administrative bodies (ZBOs), agencies New Zealand Crown entities, semi-autonomous bodies Spain Autonomous bodies with administrative functions, Sweden Boards, agencies U.K. Next Step Agencies (now called "Executive Agencies"), non-departmental public bodies Ministerial departments Departmental agencies PLAsthat implement Governance board Public law government policy The public sector administrations Differentiationo f control (PLAs) PLAsthat are guided by government policy PLAsthat independently exercise statutory powers Government Not-for-profit enterprises SOEs The NPO sector NPOS The private sector Private Small and medium enterprises enterprises Large corporations 22 Deepening Public Service Unit Reform to Improve Service Delivery Government bodies involved inpublic service delivery 2.6 Reflecting the varying legal and historical backgrounds of the countries, the organizational forms of government bodies involved in public service delivery differ widely in terms of the ways autonomy and accountability are balanced. Relatively clear are traditional ministries ("departments" in some countries) and state-owned enterprises, which define two extremes of the spectrum. In between lies a diverse range of public entities, whose Chinese counterparts are PSUs. Among them "departmental agencies" and "public law administrations (PLAs)" are two larger groups. 2.7 Departmental agencies are direct subsidiaries of ministries and function under public law or general administrative processes applying to all ministries. They are sometimes called "delegated" service providers. Their main difference from traditional, vertically-integrated ministrieddepartments i s that they are given more managerial freedom and function within a management and performance contract negotiated with the parent ministry. They do not have a governance board. The minister has formal control, but less direct than in a ministry, while the director general or chief executive has operational control. Departmental agencies are fully tax- or partly tadpartly user fee- financed. They have an annual budget and they can often carry forward surpluses. They may be subject to process controls or not. Their staffs are part of the civil service, but often with some relaxation of civil service staff processes and numbers. They usually do not have independent legal status and have no power to borrow or lend funds, but the proceeds from their assets sales may be retained. Inaddition to the UK and New Zealand, the Netherlands has also created a large number of departmental agencies. It was expected that by 2004, 80 percent of the civil service would be working in such agencies. InKorea, 23 departmental agencies have been created since 1999, which employed more than 5,000 staff and accounted for 7 percent of the government budget. 2.8 The category of PLAs itself i s a diverse group. Within any one country, a wide array of governance regimes and legal forms may be used to undertake a wide variety of functions. Across countries the legal status, control regimes, powers, size, functions, internal governance and administrative cultures vary even more. With some exceptions, PLAs are a reasonably stable well-established group, which in some cases have histories dating back to medieval times. PLAs can be significant government employers. In Germany, ministries employ only 6 percent of civilian federal employees, compared to 51 percent inPLAs.19 2.9 Incontrast to departmental agencies, PLAscan be labeled as "devolved" service providers. Although they function mostly under general rules applying to government entities, they differ from ministries in that they are separate legal entities from the state and have different governance structures. They may have a full governance board, an advisory board, or be under one-person rule. Their control i s given to the governing body, which generally delegates to a director-general or chief executive. The minister retains indirect control. They can be solely tax-financed, some may be part tadpart user fee-financed. They are often allowed to carry forward their surpluses. The status of their staff varies between full civil service controls, partial controls, or general labor laws. There are varying controls on the appointment of the director generals, depending on the l9 ofemployeesinfederalministriesworkinthedefensesector. 40% InternationalExperience in Public Service Provision and its Reform 23 status of the board. Finally, controls on borrowing, lending and retention of proceeds from assets sales vary. Three types of PLAs can be identified (Gill 2002, pp. 48-54). More details of features of these government bodies are summarized inBox 2.1. i. PUS that implement govemment policy. Examples include "EPAs (administrative public establishments)'' in France and "indirect public administrations" in Germany. In some cases, service providers such as schools and hospitals fall in this category. This type has two variants. Those of the first variant have governance boards. Control i s devolved to the board, which delegates to the director general, with indirect control from ministers. PLAs of the second variant have so-called "differentiation of the control environment," meaning that some key controls and operating rules are designed for a particular form of organization. Control i s often devolved to the director general, assisted by an advisory board under indirect control from ministers. .. 11. P U Sthat are guided by govemment policy. This type of PLA is required to be influenced or guided by ("have regard to") government policy, as opposed to complying with it. They are governed by a board. Control i s also devolved to the board which delegates to the director general, under indirect control from ministers. Examples include some New Zealand Crown entities such as universities and the National Museum, some German "corporations" such as universities. iii. PUSthatindependentlyexercisestatutorypowers. PLAsofthistypeare either governed by a board or a director-general (one-person rule). Control i s devolved to the governing body with very indirect control from ministers, as dismissal of the governing body is very difficult. Independent regulatory commissions in the U S and independent Crown entities inNew Zealand are examples. Box 2.1 Government bodies involved inpublic service delivery inOECD countries: Summary of their features 1. Ministerial Department Example: Ministry of Defense Legal Status: Public law; no separate legal identity from the state. Governance: No board, partof the civil service. Control: Minister has formal direct hierarchical control while the Director General has operational control. Finance: Tax financed; no ability to borrow or lend; typically has limited ability to carry forward surpluses. Staff Process controls on the appointment, promotion and removal of staff; input controls on the price and quality of labor. Capital Spending: Controlled separately with proceeds from assets sales not retained. 2. Departmental agencies Examples: Regulatory compliance services (e.g. issuing drivers licenses), supporting services to ministries. Country examples include: centers of responsibilities inFrance, direct federal 24 Deepening Public Service Unit Reform to Improve Service Delivery Box 2.1, continued administration in Germany, agencies inthe Netherlands, semi-autonomous bodies inNew Zealand, delegations in Sweden, executive agencies inthe UK, and PBOs inthe US. Legal Status: Public law; no separatelegal identity from the state. Governance: No board, part of the civil service. Control: Minister has formal but less direct hierarchical control while the Director-General has operational control. Finance: Fully tax or partly tadpartly user fee financed; annual budget; carry forward of surplus often allowed. Staff Generally civil servants but often with some relaxation of civil service staff processes. Capital Spending: Usually no power to borrow or lend; proceeds from assets sales may be retained. 3. PLA General Examples: Bodies responsible for funding in areas o f arts and culture, sport, sciences; bodies responsible for regulatory compliance in areas of health, safety, and environmental issues. Country examples include EPAs in France, public law ZBOs in the Netherlands, executive non-departmental public bodies in the UK and most independent agencies in the US. Legal Status: Public law; varies between separate legal body, partially separate with respect to third parties, or not separate. Governance: Varies between governance board, advisory board or corporation sole (one-person rule). Control: Devolved to governing body, which generally delegates to Director-General; minister has indirect control. 4. PLAs that implement government policy, differentiation of control Legal Status: Public law; partly or fully separate legal entity. Governance: Director-General assistedby an advisory board. Control: Devolvedto Director General, minister has indirect control. Finance: Some solely tax financed, some part tadpart user fee financed; annual budget; carry forward of surpluses allowed. Staff Partof civil service and subject to all input andprocess controls. Capital Spending: N o borrowing, lending or retention o f proceeds from assets sales. 5. PLAs that are guided by government policy Examples: Bodies responsible for funding in arts and culture, sport, sciences; legitimacy bodies that involve social partners, regulatory compliance in occupational regulation. Country example include other public interest authorities GIPs and professional EPs in France; some Crown Entities including fish and game council, universities, and the National Museum in New Zealand; corporations such as universities in Germany; some government corporations inthe US. Legal Status: Public law; completely or partially separate legalbody (with respect to thirdparties). Governance: Board. Control: Devolved to the board which delegates to the Director-General; minister has indirecl control. Finance: Annual tax revenue financed; carry forward of surplus allowed. Staff Part civil service exempt some control but within state sector framework. Capital Spending: N o borrowing or lending, but retention of proceeds from assets sales allowed. International Experience in Public Service Provision and its Reform 25 ~ Box 2.1, continued 6. PLAsthat independently exercise statutory powers Examples: Regulatory functions applying to state bodies: police complaints, human rights adjudication, and independent utility regulators. Country examples include autonomous administrative authorities in France, independent Crown Entities in New Zealand, independent regulatory commissions inthe US. Legal Status: Public law; separate legal body. Governance: Board or PLA Director-General (one-person rule). Control: Devolved to the governing body; minister has very indirect control because dismissal of the governing body i s very difficult. Finance: Levy or tax funded; annual budget; no carry forward of surplus allowed. Staff Within the civil service but exempt some civil service controls. Capital Spending: No borrowing, lending or retention o f proceeds from asset sales. 7. Government not-for-profit enterprises Examples: L a Poste inFrance, ZBOs on the way to privatization inthe Netherlands, the BBC in the UK. Legal Status: Functions under private law; separatelegal entity. Governance: Governance board, not part o f the wider civil service. Control: Devolvedto board which delegates to the chief executive, minister has indirect control. Finance: Sales revenue (and limited subsidy) financed; may carry forward surpluses, borrow and lend. Staff No staff or other input control. Capital Spending: Proceeds from asset sales and depreciation retained. 8. State-owned enterprises Examples: Air France, Air New Zealand, state-owned companies inSweden. Legal Status: Functions under private law; separate legal entity. Governance: Governance board, not part of the wider civil service. Control: Devolvedto board which delegates to the chief executive, minister has indirect control. Finance: Sales revenue (and limited subsidy) financed; may carry forward surpluses, borrow and lend. Staff N o staff or other input control. Capital Spending: Proceeds from asset sales and depreciation retained. Source: World Bank, 2004. Private sector participation 2.10 For provision of public services, private sector involvement and competitive pressures are often introduced through an approach known as "contracting out." Over the past 20 years, many OECD governments have developed a specific mandate to review public activities, both in order to contract out those which can be delivered more cost effectively on contract by the private sector and in order to create performance incentives for in-house providers as a result of the "contestability" such processes may introduce (sometimes referred to as "market testing"). The most commonly contracted-out services in OECD countries have been those which are not considered to be "core competencies" 26 Deepening Public Service Unit Reform to Improve Service Delivery of government and which are characterized by relatively high measurability and contestability: cleaning, security, catering, maintenance, printing, etc. 2.11 In spite of these important trends, a sizeable core of public services in OECD countries i s still delivered by government entities. In most OECD countries, the very large share of public funds for schools, universities and other educational institutions are directed towards public institutions. In a small number of OECD countries, notably Belgium and the Netherlands, significant public funds are transferred to private institutions. A few OECD countries are experimenting on a limited scale with giving public education funds directly to households as "vouchers" to spend inthe institutions of their choice. However, on average across OECD countries at the primaryhecondary level, only 12 percent of public funds for education institutions are spent in institutions that are privately managed. 2.12 Inthe health sector, there is more diversity. Three models have been identified by the OECD: the public integrated model, the public contract model, and the private insurance-private provider model.20 The Nordic countries, Australia, Italy, Greece and, untilthe 1990s, the UnitedKingdom, have adopted a public integrated model, financing comprehensive healthcare for all citizens from public money, with government-owned- and-operated health facilities delivering the very large majority of services. Canada, most continental European countries, Japan and now the UKbelong to the public contract category, where public payers contract with private health care providers through a single payer arrangement. By contrast, the United States has adopted a private insurance- private provider model, with the bulk of healthcare provided from private insurance and out-of-pocket payments, with only partial support from government for the old and for children of low-income families; health delivery has been through a mixture of private for-profit, private not-for-profit and government-owned facilities. Non-profit organizations 2.13 Alongside the public sector and the for-profit private sector, there can be a "third sector": non-profit organizations (NPOs). A NPO is an organization that i s barred from distributing its net earnings, if any, to individuals who exercise control over it (Hansmann, 1980), a constraint known as the "non-distribution constraint." The value of the NPO as an institution i s primarily about trust and motivation. By institutionalizing a weak profit motive, NPOs help build up trust of consumers/donors toward service providers. Such trust reduces transaction costs and makes the market work for some transactions that would otherwise be unattainable. In addition to the non-distribution constraint, NPOs are often run by faith-based individuals, which can be helpful in establishing trust and reducing transaction costs. In terms of financing, the NPO represents an alternative to the government and the for-profit sector when it finances services with donations. The non-distribution constraint implies a donation from the founders of a "0, who could have reaped a financial return had they invested inthe for- profit sector. In addition, with its institutional advantages in addressing contract failure, 2o Health Care Systems: Lessons from the Reform ExDerience, OECD. Cited inPilichowski, 2004. International Exverience in Public Service Provision and its Reform 27 NPOs are often able to finance a significant portion of the services they provide with donations.21 2.14 Insome countries, NPOs can provide a largepart of public services-for instance in the Netherlands, where a majority of schools and a large part of medical services are provided by these entities (albeit with the bulk of funding from taxation through public budgets). While the NPO as an institution has demonstrated itself as an alternative to both the government and the for-profit sector, it i s important to realize that the aggregate activity o f the NPO sector pales by comparison to that of the government. Developed countries differ greatly in the extent to which various services are provided by the NPO sector, but the fair description seems that it has a niche role, reaching some constituencies not adequately served, but not substitutingfor the government (Schick 2004, p. 15). PUBLICSERVICE REFORMINOECDCOUNTRIES 2.15 In recent decades, public sector organization and management has undergone major reforms in several OECD countries, in particular New Zealand, the UK, the Netherlands and Sweden. The New Zealand model went farthest in introducing market- like mechanisms in the public sector, including the introduction of separation of purchaser and provider, establishment of a public services commission to manage performance contracts, and major reforms in budget management and accounting to better reflect true costs of policies. To a lesser degree, these techniques found their way into service provision inother countries, although countries such as Germany and France stayed much closer to the traditional bureaucratic model. These organizational changes within the public sector went hand-in-hand with a refocusing of the state by means of privatization and divestiture of tasks that couldbe better left to the market. The case of New Zealand 2.16 The New Zealand model relies on a web of personal and organizational contracts. One key contract i s a "purchase agreement" that i s negotiated between a chief executive and the minister responsible for the department. It specifies the resources to be made available during the year and the outputs to be produced. A "performance agreement" between the State Services Commission (representing the government) and each departmental chief executive specifies the key results expected during the year. There also are individual employment contracts for senior and middle managers, "statements of intent" prepared each year by Crown entities, and other contract-like arrangements. 2.17 The main problem with these contracts i s that although the government's purchasers (ministers) and providers (departments and Crown entities) may be organizationally and legally separate, they are locked into an interdependent relationship. For example, ifthe Ministry of Education must rely on public schools for teaching children, writing a contract that sets the price and output targets for schools will not do much to improve either performance or the government's capacity to influence educational results. In fact, ministers still purchase most services from their own 21 It i s worth noting, however, that the nonprofit sector i s not primarily financed by private donations. A study found that on average, private giving accountedfor 10 percent of revenues of the nonprofit sector in seven developedcountries. See Rose-Ackerman, 1996,pp.702,707. 28 Deepening Public Service Unit Reform to Improve Service Delivery departments. In contrast to contracts executed in real markets, government has little recourse for failure to perform. It does not gain much leverage by having the option of canceling the contract or stopping payment. The prospect of firing public employees may do more damage to the political interests of ministers than to the department that employs them. Duringnegotiations, ministers typically know much less about the outputs they are buying than managers know about what they are selling. In many New Zealand departments, the chief executive drafts the purchase agreement and presents it to the minister, who signs it after making a few marginal adjustments, such as adding a minor task to the list of outputs. 2.18 New Zealand has invested in independent assessments of its reforms.22 The State Services Commission, which has had a leading role in managing the new system, has been forthright in criticizing the contract model that underlies purchase agreements. "The model is that of the marketplace: buyer and seller haggling over price against a backdrop o f competitive supply, discriminating demand, efficient price, symmetrical information and minimum transaction costs. Experience has shown, however, that none of the model's conditions can be reliably satisfied, even in a weak form, in the New Zealand Public Service. The market for New Zealand Public Service inputs i s characterized by monopoly supply, compliant demand, and arbitrary price, asymmetry of information and time horizons, and significant transaction costs, many of which arise well after the transaction. The concept of an efficient market fundamentally misrepresents the nature of New Zealand public service's production and exchange.'' The case of theUnitedKingdom 2.19 The United Kingdom has probably invested more care in establishing "agencies" than any other country. Much of the UK reform effort has focused on better performance management. 2.20 Inwhat is known as the "Next Steps" process, which has been replicatedinfew other countries, agencies go through a life cycle that typically consists o f eight steps: The government identifies candidate agencies. The Treasury or Cabinet Office (which has government-wide responsibility for public management) reviews options for carrying out the agency's missions. This "prior options" review may conclude that the activity should be privatized or kept within a department. The government examines the relationship of the proposed agency and the department to which it i s responsible. The key issue here i s whether the department i s adequately preparedto supervise the agency. The agency and government negotiate a framework document that specifies the matters on which the agency may exercise discretion, 22 In2001 the government establishedthe Review of the CentreAdvisory Committee to review the state sector. The report and recommendationsof the Advisory Committee were releasedinDecember 2001. A briefing paper to the Minister of StateServices on the Integrated ServicesDelivery Programme was published on July 20,2004. See: http://www.ssc.govt,nz/display/document.asp?NavID=105&DocID=4050. International Experience in Public Service Provision and its Reform 29 reporting requirements and other operational matters. The framework document serves as the legal or contractual basis for each agency. (v) The government recruits a chief executive, often from outside the civil service, and gives the person a fixed-term contract. The contract specifies performance expectations of the chief executive. (vi) The government decides each agency's budget annually, along with performance targets for the year. This entails much discussion and negotiation between the parent department and agency. (vii) Each agency reports annually on its finances and performance targets. In some years, the government publishes performance results against these targets. (viii) Periodically (every 3-6 years) each agency i s evaluated and the government reconsiders its status and framework. 2.21 Behind this sequence of activities i s the idea that an agency must be capable of self-management before it i s entrusted with public money and responsibility. When the Labour Government took office in 1997, it continued the agencies, but gave much more priority to the policy andcoordination role of parent departments andcentral agencies. In effect, it subordinated agencies and the work they perform to the functions and responsibilities of departments. Periodic evaluations are required to take account of the fact that "agencies do not exist in isolation from the rest of government, and so greater emphasis will be placed on considering the performance of agencies in the context of their contributions to wider Departmental and Governmental objective^."^^ 2.22 Many o f these objectives have been articulated in an important innovation of the Labour Government in expenditure review. The government's biannual "Spending Review" process has been expanded into the means by which departments negotiate formal, published agreements on exactly what they will accomplish in the next three years. These Public Service Agreements (PSAs), which are agreed between the Treasury and the relevant departments, link resources and results by specifying the targets each department commits itself to achieve. For example, the PSAs for 2001-2004, set out approximately 160 specific performance milestones. These are regarded by the government as "transparent commitments which the Government can be held accountable for, and all departments will report on every year."24 As in other recent British managerial innovations, PSAs have redefined the role of agencies in terms of contributing to departmental-rather than their own-performance. PUBLIC SERVICE REFORMINTRANSITION COUNTRIES 2.23 At the start of their public service reform, transition countries typically had much larger public sectors than OECD countries. In many cases, the public sector covered virtually everything, including administrative functions, "productive" functions 23 ExecutiveAgencies 1999Report, Chapter 2, p.10. See http://www.archive.official-documents.co.u~documen~cm46/4658/chapter2.pdf. 24 The UK Treasury, Public Service Agreements, 2001-2004. See http://www.hm- treasury.gov.uk./Documents/Public~Spendingand_Services/Public~Service~A~eements~2OO 1_2004/pss- psaJntro.cfm. 30 Deepening Public Service Unit Reform to Improve Service Delivery and "non-productive" functions. Whereas the reform challenge for OECD countries has been largely to do with balancing autonomy and accountability, transition country experience has involved a much more fundamental redefinition of the boundaries of the state, a process of decentralization to lower tiers of government and the creation of institutions to support a mixedeconomy. Financing 2.24 Service delivery has suffered greatly from financial difficulties in the 1990s in most transition countries. The sharp fall of GDP was the first factor behind the difficulties. Among 24 transition countries, only 5 saw their 2000 real GDP recovered to the 1990 level (World Bank 2002b). The share of GDP devoted to key public services fell as well, but to a varying degree across countries. Decentralization was another factor that contributed to the financial difficulties. The past decade has witnessed a dramatic decentralization in many transition countries. Many social services previously provided by central government, or else previously provided by enterprises to their employees, have been transferredto sub-national tiers of government. 2.25 This transformation has not been a smooth one in many transition countries. The divestiture of social services has frequently been done in a way that leads to confusion over which level of government i s responsible for what. For example, in the health sector in Slovakia, the ongoing confusion regarding the respective responsibilities of central and local government has undermined service delivery. In addition, many transition countries have shifted social responsibilities to local governments, but have not ensured that local governments have enough resources to manage these responsibilities, either through bigger transfers or greater tax autonomy. Regional disparities in income have also led to disparities in terms of provision of government services such as education. 2.26 Severe difficulties in public financing have compromised the quality of services in some transition countries. InGeorgia, 43 percent of primary and secondary schools in urban areas got textbooks for all children, compared with 27 percent inpoorer rural areas. In Russia, data for 1997 suggests that 41 percent of all patients could not afford to purchase required drugs and 11percent could not afford any kind of medical treatment (World Bank 2000b). Similarly, 37 percent of pregnant women inTajikistan didnot seek prenatal care because it was unaffordable. As a recent World Bank report concluded, "in several countries, a privately financed, unregulated system of health care i s in a public shell" (World Bank 2002b). 2.27 In the absence of robust markets, strong governance institutions and effective regulation, fees for essential public services are hidden taxes, and highly regressive ones which widen the gap between rich and poor. In some transition countries, the combination of inadequate funding, poorly designed user charging and weak institutions for regulation, transparency and accountability have damaged public service ethics and led to an increase in administrative corruption in social service delivery. In education, parents are asked to supplement the salaries of education personnel through under-the- table tutoring schemes-if they want their children to pass. Similarly in health, inadequate funding and weak institutions for regulation, transparency and accountability have ledto under-the-table payments for drugs and other treatments in short supply. Intemational Experience in Public Service Provision and its Reform 31 Provision 2.28 Many transition countries have embarked on privatization programs which have included large service-providing entities. Some countries-notably, Russia, Latvia, and Kazakhstan-have attempted to establish "functional review" mechanisms in order to redefine the boundaries of the state by abolishing, rationalizing, privatizing or transferring functions of government, and by restructuring and reforming remaining functions. Overall evidence i s mixed as to how such processes work in practice. However, in some particular cases, there have been successes. In the health sector, for example, Estonia has achieved positive results in developing its primary health care (PHC) system by changing doctors from salaried employees to independent contractors paid through capitation (a form of privatization). Evaluation found both allocative efficiency and technical efficiency improved (Koppel, A., K. Meiesaar, et al, 2003). From the perspective of patients, it has also been found that most of the Estonian population has accepted the new system and the satisfaction rate has increased. Family doctors are in the center of the primary health care systems, with their professional competence playing a great role in making the system acceptable for the patients (Polluste, Kalda and Lember, 2004). 2.29 Governance of public providers has experienced radical changes in some countries as well. Most notably, Latvia and Kazakhstan have created a significant number of arm's-length agencies as part of their transitions. Many parts of government were restructuredinto a series of arm's-length agencies with a contractual relationship to their sponsor ministry as a way o f introducing market-style incentives and private sector- style efficiency and escaping bureaucratic rigidities. Such agencies were established as non-profit state stockholding companies. They were financed partly from the state budget and partly from their own incomes, raised either by charging fees for services or levying taxes. However, during the course of the decade it became clear that premature marketization was fostering corruption and undermining the state. Government has now recognized that the establishment of these agencies was not accompanied by adequate institutions for regulation, transparency and accountability. Funds were increasingly kept outside the Treasury system, undermining transparency and meaning that the State Treasury's control over such institutions was weak to nonexistent. Since the late 199Os, the Latvian government has been attempting to re-regulate these agencies. A process has started to transform state non-profit companies into public agencies on the basis of three new laws. PUBLIC SERVICE REFORMINLATIN AMERICA AND THE CARIBBEAN2' 2.30 The last two decades have witnessed radical reform in public service delivery in most countries in the Latin America and Caribbean region. There have been both successes and failures. Pronounced improvement has been made in coverage o f basic services with increased public spending. Establishing compacts with agencies or actors outside the mainstream state apparatus has, overall, produced good results at least in the short run. Giving individuals choice over whom they get their services from has, in the limited areas in which it has been implemented, generated powerful pressures for results 25 This section draws on World Bank, 2004. 32 Deepening Public Service Unit Reform to Improve Service Delivery on front-line providers. Client participation inthe management and operation of services has also shown good results in the case of services that allow a relatively small production scale. However, large inequalities in access to services remain a feature in most countries in the region, and the fundamental challenge of reforming the traditional state institutions i s still to be coped with. Progress incoverage of basic services 2.31 Coverage of primary and secondary education increased significantly between 1980 and 2000 in the region. Considering the 20 largest countries, average primary enrollments went from 81 percent of the relevant age group to 92 percent, while the corresponding figures were 40 and 58 percent for secondary education. Coverage of health services also increased in a significant manner during this period. This was largely achieved through a substantial increase of public resources. Public spending on education for the 20 largest countries inthe region grew from 3.3 percent of GDP in 1975 to 4.2 percent in 1999. The increase in health expenditures i s even more marked. Several countries have experienced large increases in public spending (above 50 percent) in these two services. The fact that, in most cases, the expansion in coverage of social services was done fundamentally by increasing budgets suggests the presence of large inefficiencies which, in some cases, may question the sustainability of the overall efforts. On quality of services, there are clear signals that suggest both poor average quality of services and significant differences among social groups. Establishment of compacts with non-governmentproviders 2.32 Overall, government bureaucracies in the region remain generally weak, ineffective and largely unaccountable for results. State modernization reforms have upgraded governments' institutional capacities considerably, but very few countries in the region have been able to move decisively toward developing a results-oriented public administration. Governments in the region have sought alternative, often innovative, ways to avoid the deep difficulties of reforming bureaucracies (central and local) as a means of improving service delivery outcomes. Many of these alternative ways involve a combination o f contracting out services to private firms and non-government organizations (NGOs) and relying on alternative bureaucratic structures that bypass more traditional ones that are hard to reform. Overall, these efforts produced positive results in terms of improving service delivery outcomes. There are many successful, and mostly well-documented, examples of such initiatives (see Box 2.2 for two of them). At the same time, far from being a silver bullet for service delivery reform, bypassing state bureaucracies also raises a series of broader, mostly long-term, common concerns. Effective compacts for service delivery (within or outside the state) require the existence of basic commitment and support of the policymakers at a minimuminterms of initiating and sustaining them. When public bureaucracies do not work well, short-cutting them by contracting out services to private firms and NGOs or setting up parallel, independent, agencies in charge of service delivery i s showing some good results in the short run. In the longer run, however, both their sustainability and full effectiveness depend on whether the rest of the state apparatus improves its capacity and becomes more performance-oriented. International Experience in Public Service Provision and its Reform 33 Box 2.2 Establishingcompact with non-government service providers: two successful examples The Bogotri school concessions program (which created 18,000 new school places) i s an important example o f how private management of public schools could work if used more extensively. The municipal education department constructs new, state-of-art, schools in low income areas each able to accommodate between 800 and 1,200 students. The opportunity to manage the school is offered ina public procurement process, where bidders (good quality private educational institutions) are evaluated on their proposed management plans. Concessionaires are obliged to provide educational services to poor children, and in return receive remuneration from the department on a per-student basis. While the contract establishes clear standards that must be upheld, the concessionaire has full autonomy over school management and i s evaluated solely on results. Concessionaires have already produced striking results in 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 Alegria schools are the largest case of contracting out in the education sector. They operate in 14 Latin American countries covering approximately one million students. Evaluation of the program's operations innine countries indicates that schools integrated into Fe y Alegria outperform traditional public schools, with lower repetition, dropout, and higher overall rates of progress in school and higher overall retention rates on average. Critical to Fe y Alegria'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 o f learning styles. The dynamic at play inthis case i s strongly associated with the altruistic and "empowerment"-oriented approach that characterizes Fe y Alegria. In other words, the effects of contracting out appear to be closely related to the nature of the contractor. Source: WorldBank, 2004. Empowerment of clients with choice 2.33 The introductionof choice as a mechanism to promote client power over service providers has been a powerful one, albeit of restricted coverage. Inthe case of education services, the two key experiences with choice are those of Colombia and Chile. Colombiaintroduced inthe first half of the 1990sa secondary education targeted voucher program which involved approximately one quarter of the municipalities in the country and close to 2,000 private schools with the main objective of increasing the transition from primary to secondary education for poor students. Chile, on the other hand, represents a unique case in the sense that a nationwide system of "quasi-vouchers" directed to all municipal schools and non-fee-charging private schools, was established with the main objective of promoting competition among schools. In the health sector, once again, Chile and Colombia introduced paradigmatic reforms that put competition and choice at the center. Overall, the results of competition have been very positive in all sectors where it has been tried. This, nonetheless, should not obscure the fact that the conditions for choice to work are not minor. The proper functioning of a system based on choice and competition depends on the presence of active public policies to compensate both for informationalasymmetries and differences inpurchasing power and needs. Without such compensatory actions, increased social stratification becomes a significant danger. The presence of independent providers operating in a competitive environment is, naturally, key as well. 34 Deepening Public Service Unit Reform to Improve Service Delivery Clientparticipation 2.34 Client participation in the management of service delivery i s the second way in which clients are empowered. Overall, evidence in the Latin American and Caribbean region suggests that direct client involvement in the management of service delivery can yield positive results when the services involved allow a relatively small production scale and rather simple management challenges. The experience of community-run schools in El Salvador and Nicaragua, where each school is required to establish an elected council composed only of parents in El Salvador and of parents, teachers and the school director inNicaragua, provides solid evidence on the benefit of a participation approach to client power. The key aspect of this model i s the control parents have over teacher management decisions, which i s exactly the aspect that has proven hard to reform in a state-managed school system. In contrast, evidence of effective experiments with community-run health centers i s rather limited. There are a number o f factors that seem to conspire against community participation in health centers, compared to that in school management. Firstly, in most cases users of health services have much less frequent relationship with providers than parents have with schools, making it more difficult to solve the collective action problem. Secondly, informational asymmetries tend to be very large in the case of health services, increasing the opportunities for health professionals to capture control. LESSONSCHINA FOR 2.35 What does international experience suggest for China? Although international experience does not provide any prescription in developing an overall strategy for PSU reform, the following may deserve some consideration. 2.36 First, public service reform involves modernizing the state as both service provider and financier. The traditional role of "watchman" notwithstanding, the definitive role of the modem state i s as service provider and financier (Schick 2002, p. 44). A comparison of experiences of the three groups of countries discussed in this chapter suggests the critical role the state plays in financing services. While it i s surely not sufficient, adequate financing i s the necessary condition for any system to achieve desirable results. It has long been recognized that a strong state i s essential in underpinning markets by protecting property rights, enforcing contracts, providing legal due process, and establishing conditions that encourage capital flows and other transactions. But the contribution of a modem state to economic growth goes beyond that. Inensuring equitable and adequate delivery of services such as education and health care through its role of service financier, the state empowers individuals by bequeathing them with rights and sheltering them against critical adversities. This compares with the market, which empowers individuals by giving them choice and opportunity. Ultimately, it has to be empoweredindividuals who drive innovation and growth. 2.37 Second, diversity and flexibility of organizational forms matter. Instead of searching for the "best" organizational form, or the "best" degree of autonomy for all service providers, maintaining adaptability and flexibility may matter more. This is perhaps a lesson that can be drawn from the sharp contrast between the organizational forms of public providers in China and OECD countries. In China, almost all public providers are PSUs if they are neither core government departments nor SOEs. In Intemational Experience in Public Service Provision and its Reform 35 contrast, there i s a vastly diverse range of organizational forms between core government departments and SOEs in the OECD context, characterized by varying degrees of autonomy and ways of achieving accountability. Arguably, this makes them easier to adapt to the varying nature of services they provide as well as the environment surrounding them. 2.38 Third, compacts can be apowerful instrument inimproving service delivery, but sound performance management i s essential. All the three groups of countries covered in this chapter tried to employ compacts with service providers as an instrument to strengthen their incentives, with varying results. The successes achieved in Latin America and the Caribbean as well as in countries like Estonia highlight the potential positive effect of a greater reliance on compacts in dealing with non-state service providers, when certain conditions are met. On the other hand, the New Zealand model demonstrates its limitation when it i s applied within the state sector in an extreme manner. Ineither case, however, the use of compacts is the means and not the end. What matters for the degree of success i s the soundness of performance management. The use of compacts i s only one factor that i s necessary for sound performance management. 2.39 Fourth, empowering clients can be rewarding. While the government may motivate and discipline service providers through compacts, clients can also do so through their interaction with service providers. A critical element for good results i s empowerment of clients through two ways: increasing their choice among service providers and enabling their participation in management of service delivery. The experience of some Latin American and Caribbean countries shows the potentially high reward of this strategy. 2.40 Fifth,there is no silver bullet. There are many ways to improve service delivery results, but their effects are often highly "circumstance dependent." In addition to the varying results brought about by the greater use of compacts in different countries, the effects of client participation in the management of schools and health centers in some Latin American and Caribbean countries provides another example. The implication i s that public service reform must be designed with full appreciation of the circumstances surrounding a particular service in a particular time and place. 3. TOWARDSA PSUREFORMSTRATEGY 3.1 To improve public services, China needs to embark on a comprehensive and coordinated reform process for PSUs. In light of China's experience, as well as that in OECD countries, this process i s likely to be long and incremental, rather than one "big bang" reform done according to a blueprint. Also, given the enormous diversity of the services provided by PSUs and the various stages of reform in the different sectors and localities in China, one size i s unlikely to fit all PSUs, and the reforms will likely be sector- and locality-specific. Experimentation in reforms will pay off: the theories and practices on public service delivery are so wide-ranging, evolving, and far from uniform across countries. It i s therefore hard to predict precisely what would work in China, and it is appropriate for any large scale implementation of key measures to be preceded by experimentation in a variety of localities and sectors. 3.2 Nevertheless, there do exist common elements in the reforms that can improve the quality and efficiency of service delivery, and China's past reform experience suggests that a well-thought-through overall strategy and a guiding framework established by the central authorities are critical to the success of regional and sectoral initiatives. The recommendations offered below are intended to form key elements of such an overall strategy for PSU reform in the next stage. They aim fundamentally to change the incentive system for PSUs and their supervisory departments and secure an adequate supply of resources for improvements in service delivery results. Taken together, the reforms should lead to a new system of public service delivery commensurate with the ambitions of axiaokang society. REDEFINETHE ROLE OF GOVERNMENT 3.3 While China has made great achievements in the transition to a socialist market economy, a reconsideration of the role of the state in the service sector i s warranted. Such a redefinition i s inevitably the central theme of PSU reform, and a necessary step toward a "service-oriented government." 3.4 Generally speaking, the government needs to take three basic decisions in redefining its role in service delivery. First, it needs to decide on the scope and priority of intervention. Not all services need government involvement in delivery. Indeed, many of the activities performed by PSUs in China today could be considered commercial production of private goods or services, which can best be left to the non- state sector and the market. Government intervention in service delivery i s justified by failure of the private market in generating satisfactory results in terms of both equity and efficiency. In addition to the well-known inability of the market to achieve equity, there are four common cases where the market may fail to deliver the most efficient results as well: they are public goods, externalities, natural monopoly and asymmetric information. The desirable scope of government intervention in service delivery needs to be 37 38 Deepening Public Service Unit Reform to Improve Service Delivery determined through careful review of market failure in the delivery o f each service involved and in light of the availability of public resources. 3.5 When the government figures out the scope of its involvement in service delivery, it needs to go further to decide on a division of labor among various levels of the government: should the intervention be done by the central government, the provincial government, or the municipal/county government? For each level of the government, the third decision would be about how to intervene. It should consider whether it wants to regulate non-state production of that good or service, or consider provision by the state. If the decision i s to provide, it should consider whether it should just fund the activity, or also produce it. If the government decides to fund it, it should further decide how-contracting an enterprise or non-profit organization to do it, or funding individual consumers such as in the case of vouchers. Finally, the government can produce the service itself, and it needs to decide what type of organization in the public sector i s best suited for what kindof services. 3.6 In this context, it is important to realize that the three instruments of government intervention in service delivery, i.e., regulation, financing and direct provision, are designed to address different kinds of market failure, and justification for one does not extend automatically to another. Many services do not require government involvement much more than for sound regulation. Some services require government funding, but not necessarily provision by public sector providers, as they can be contracted out. Yet other services may need to be provided by public sector providers, butcan be funded at least partially by chargingusers. Inessence, the responsibility of the government i s to ensure that services are delivered in a satisfactory manner, not necessarily to deliver services itself. 3.7 In China, the most frequently usedcriterion for continuing government funding and state ownership of the PSU i s the notion of "gongyi Xing (public benefit relevance)." Public benefit relevance i s certainly a necessary threshold when the government i s considering the scope and priority of intervention. Pure market failure does not justify government intervention when there i s no concern of public benefit. However, without full consideration of market failures in terms of both equity and efficiency, "public benefit" as a concept would be too vague to be usefully applied inpractice. Inparticular, it i s hard to justify the call for public financing, public ownership and participation of not-for-profit entities in delivery of services of "public benefit" without referring to market failures. Furthermore, "public benefit relevance'' has no role to play in determining how the government should use the three instruments of intervention. For example, one may have good reason to regard both medical services and pharmaceutical production as equally "public benefit related," but it would be wrong to argue that government intervention in both fields should be done in the same way in terms of financing and provision. A school and a kitchen supplying lunch to students of the school are both service providers, and there i s good reason to argue that both are "public benefit related," but this tells nothing about what the government should do in financing and providing both services. What really matters in determining the role of the government in service delivery i s equity consideration, as well as the four cases of market failure: public goods, externality, natural monopoly, and asymmetric information. Towards a PSU Reform Stratenv 39 IMPROVE FINANCES 3.8 China's government spending on social services, despite recovery in recent years, does not seem sustainable, if "people-centered" development i s to be achieved. It seems quite possible that the reform when taken as a whole will involve more government spending on public services, despite possible cost savings when the government withdraws from certain areas. The very large majority o f PSUs across government as a whole are involved in provision of basic education and primary health and operate in rural areas at the county and township levels; given the rising trend of income inequality in China, these are likely to need more government funding if such services are to become more accessible for the poor and those in rural areas and to reach an adequate minimum standard of service. It i s very likely that China's development path will have to be characterized by an increasing role of the government in financing key public services, as was the case in OECD countries in the past century. The temptation o f shifting financing responsibility to "the society" may need to be carefully resisted. 3.9 However, this does not necessarily imply a need for a higher share of government spending in GDP. China's government spending, with some 22 percent of GDP, i s not low by East Asian standards. What the reformmustfocus on is prioritization and refocusing. Under the current system, scarce public resources are still funding activities that are not justified by either equity or efficiency considerations. At the same time, services considered as key to achieving national goals are underfunded. Under- funding is often the most pressing in the poorer regions of the country. This situation calls for better allocation of public resources over services, and across the country. This requires the government to take actions inthree areas. 3.10 Devolving some services to private sector or SOEs. Firstly, the government could identify services that can best be funded and provided on a commercial basis, and devolve them to the private sector, or if ownership i s a concern, to SOEs with a profit motive. Such services as engineering and consulting services now often performed by research institutes under government could be left to commercial for-profit entities, while state funding could be limitedto contract work. 3.11 Reforming budget management to better align resources with priorities. Secondly, the government could reform budget management so as to better align budget resources with the priorities of government. While the central government level has made major progress in budget reforms in recent years, many local governments still follow a largely incremental process that does not allow retrenchment from areas that are no longer a priority. To carry forward the reform, the government could consider developing a medium-term expenditure framework; taking decisions on tax policy together with decisions on spending; requiring detailed budget proposals that establish the link between proposed spending and China's priorities; and developing capacity for evaluating expenditures in terms of their impact on these priorities. Information collection on PSUs and transparency should be improved to provide the basis for monitoring and oversight of their operations. 3.12 Realigning intergovernmentalfiscal relations. Thirdly, the government could realign intergovernmental fiscal relations to create the preconditions for more equitable 40 Deepening Public Service Unit Reform to Improve Service Delivery access to services. This would require a more explicit assignment of functions to each level of government and more equitable distribution of fiscal resources among sub- national governments (see World Bank 2002a for some specific recommendations). Reform o f local government finance i s now receiving urgent attention from the top leadership. InMarch 2004 the State Council appointed a special committee to investigate the fiscal problems of county and township governments. Going forward, the intergovernmental fiscal arrangements will have to be fundamentally reformed, including the possibility that responsibilities for the most vital services such as basic education and public health may have to be reassigned to higher levels, as was done for rural education in2001.~~ REORGANIZETHE PsuSECTOR 3.13 The choice of organizational form for a service provider (public or private, for- profit, not-for-profit, autonomous entity, departmental agency, department) depends on a host of factors, including the nature of the service produced, the measurability of the outputs and outcomes, the degree of contestability of the market, and the desired extent of control over the activity of a service provider. A service for which outputs are easily specified and measured can be more easily contracted out to an enterprise or devolved to an autonomous agency. Services for which this i s not the case can be better provided within government itself, with public sector ethics and input control as means to ensure the service. But for some services, notably those of independent regulation of markets, an arm's-length relationship i s desirable for reasons other than efficiency o f service provision. Efficiency gains from more independence or contracting out are more likely to arise if there i s competition, either "on the market," or "for the market." 3.14 China faces considerable constraints in its choices of service providers. The NPO community i s still relatively small and is unlikely to play a large role in service delivery any time soon. PSUs have gained a substantial degree of autonomy over the past decades, making it likely that additional managerial autonomy without fundamental strengthening of accountability may have limited immediate results in terms of improved efficiency. And some services that are supplied by the market in other countries cannot be so inChina because the legal andjudicial framework and institutional capacity needed to enforce complex contracts i s not available. Determining the appropriate organization for providers of each service in a particular region will require detailed review of the particular circumstances, but China i s likely to benefit from the following reform actions. 3.15 Creating two categories of organizationalforms for PSUs. First, a legal distinction between devolved and delegated PSUs could be created for existing PSUs. 0 A delegated PSU could be defined as one that is directly affiliated with a core government department without independent legal status and accounting system, and i s governed by a manager appointed and supervised by the head of the government department. The manager has relatively limited autonomy in operation, which i s larger than a core government department and smaller than 26 InJuly 2001 the State Council mandatedthat the responsibility for paying rural teachers would move to the county level, replacing the townships. Towards a PSU Reform Strategy 41 devolved PSUs. Monitoring i s more of the nature of process monitoring than results monitoring. Personnel systems of delegated PSUs couldbe integratedwith that of core departments and put under the management o f Post and Establishment Offices and ministrybureaus 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 those of core government departments. No delegated PSU should be allowed to borrow. This internal governance structure may fit most law and regulation enforcement PSUs except for a few independent regulators, and perhaps many other PSUs that provide services directly to government departments, such as policy research, information technology, data collection and management, the development and management of training programs for staff, etc. e A devolved PSU, on the other hand, could 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 objectives have been achieved. (ii) Some devolved PSUs could be governed by a board appointed by government departments that act as their owners, or those that act as policymakers, with proper participation of consumers and representatives o f other stakeholders. (iii) Appointments and dismissal of managers could be determined by the board wherever it i s created. Other hiring and firing decisions could be retained by the board (e.g., in the case of key technical and managerial staff) or delegated to the management. The board and the management should be given the autonomy to establish a human resource management system for their staff that they believe will work better in their unit. (iv) Revenue generation of devolved PSUs should be allowed, but subject to strict rules of accounting, auditing and reporting. 3.16 Establishinga legal and regulatoryframework to allow for a broader variety of providers. Second, a legal and regulatory framework could be established for a broader variety of public service providers, including public sector organizations, to formalize a larger range of organizational forms. There is no single pattern of organizational form that i s the "best" for China's service providers. Economic theories and international experiences may suggest a range o f ownership and governance structures, each of which has its own merits and therefore deserves consideration, but have little to say about the "best" combination of them. What an overall strategy for PSU reform can do i s to adopt a pragmatic approach and open the field for competition: give equal opportunity to all feasible organizational forms to present their strengths and weaknesses in all sectors and regions, and let consumers and policymakers at each level of the government make the selection. In light of international experience, China could consider seven feasible organizational forms for service providers, as shown inTable 3.1, Under government ownership, there can be four possible organizational forms ranging with increasing autonomy: core government departments, delegated government PSUs, devolved government PSUs, and SOEs. Outside the public sector, service providers can be either NPOs or for-profit private enterprises. NPOs, which are now called "min-ban non-enterprise units," could be renamed as "min-ban PSUs," while existing PSUs can be properly called "government PSUs." Private enterprises can be either small andmedium 42 DeeDeninn Public Service Unit Reform to Improve Service Delivery firms in the legal forms of sole proprietorship and partnership (for example, private clinics and schools owned by individual health and teaching professionals), or joint stock corporations. Some combinations of these organizational forms can also be useful, such as publicly-owned privately-run service providers regulatedby the government. Table 3.1 Feasible Organizational Forms of Service Providers in China The The Public Sector Nonprofit The Private Sector Sector Small and Core Delegated Devolved medium Large government (zhishu) (duZi) SOEs " 0 s or min- enterprises inthe enterprises in departments government government bun PSUs forms o f sole the form of PSUs PSUs proprietorships, corporations partnerships 3.17 T o establish such a legal and regulatory framework, there may be a need to draft at least two new regulations: one setting up rules governing delegated and devolved government PSUs; another to define the NPO and set up rules governing its operation and regulation. Given the existing Civil Codes that specify four types o f legal persons, some amendments of the law may be required to create a legal basis for the drafting of these regulations. 3.18 To ensure efficiency improvements in the longer run, China should consider a more active promotion of NPOs that could over time compete in service delivery. Changing tax laws to promote private financing for NPOs and establishing accreditation mechanisms that would allow "0s to become certified to receive public funding to deliver public services would be steps inthat direction. Government should also invest in managerial talent among the civil service ranks, so as to allow for more autonomy within the public sector over time. 3.19 Broadening the participation of private providers. Third, the government could further broadenthe options for the private supply of public services, including from enterprises as well as NPOs. This would require in the first place a strict enforcement of the Government Procurement Law and a further review of it in a longer term. Capacity building is critical to get government departments prepared to handle contracting and monitoring in an effective manner. 3.20 Reorganizing PSUs through split and merger. Fourth, PSUs could be reorganizedthrough split and merger. Where possible, diverse service delivery functions could be separated into multiple PSUs to achieve better focused organizations. Commercial functions could be separated into PSUs that are slated for divestment and commercialization. Small PSUs at township levels could be considered for merger to achieve economies of scale, for instance inagricultural extension. STRENGTHENACCOUNTABILITY 3.21 The increased autonomy that many PSUs have gained in the last decades, not least because of the diminishing share of budgetary funding, needs to be matched by an increase in accountability over the use of those resources and the performance achieved with those resources. Accountability of service providers i s usually achieved through Towardsa PSU Reform Strategy 43 two routes, as shown in a framework of accountability relationships in the World Development Report 2004 [Figure 3.13 (see World Bank 2003). The first route involves the interaction of service providers with their clients in the marketplace. Adequately empowered clients may hold service providers accountable to a large extent, often through choice among competing providers and participation in the management of service delivery. The second route has the government playing a key role in holding service providers accountable, and in the meantime being held accountable by citizens. Between the government and service providers, compacts (explicit contract and implicit agreement) are the primary means o f accountability; while accountability of the government itself i s achieved through the exercise of citizens' voices. For both routes, internal governance of service provider organizations i s crucial for accountability mechanisms to function well. In view o f China's particular situation, PSUreform could aim at substantial improvement in accountability through the following reform actions. Figure3.1 A Framework of Accountability Relationships inService Delivery accountability Short route Services 3.22 Establishing and enforcing performance management compacts. First, better accountability can be achieved by establishing and enforcing performance management compacts with ministries, local governments and PSUs. Within the government, the central government could establish service standards for each ministry, sector, local government, and individual PSU. This includes specifying the role of each level of government in public services, and minimum service standards for those functions. In addition to the realignment of intergovernmental fiscal relations, the division of decision- making power among various levels o f the government could also be reviewed to ensure that service delivery responsibilities are matched with adequate resources and decision- making power. Between PSUs and the government departments that are responsible for monitoring the performance of PSUs, resource agreements could be established to specify performance standards and rules on the use of budgetary and self-generated funds, including budgeting and treasury management. Government capacity building on monitoring and evaluation is needed to better assess performance against these standards and put the right incentives in place for better performance. Mechanisms include inspections and associated eligibility for public funding, as well as monitoring by the public through service scorecard surveys, league tables andthe like. 44 Deepening Public Service Unit Reform to Improve Service Delivery 3.23 Strengthening financial accountability. Second, strengthening financial accountability i s also crucial. At the institutional level, one basic step towards greater financial accountability i s to gradually implement a concept of complete cost accounting inPSUs. This involves inclusionof all inputsof public resources, most notably, landand fixed assets, subsidies, tax preferences, and intangible assets, in cost accounting and reporting to produce a clear idea of exactly how much public resource has been invested in a particular PSUin a certain period. The management system for "administrative and PSU (xingzhen shiye Xing)" state assets needs a major reform to set up a functioning framework.27 User charges on key services such as education and health should be limited. No delegated PSU should be granted the right to borrow from banks. For devolved PSUs, all those that demand the autonomy to borrow must be prepared for the potential risk, including liquidation and closure. Those PSUs for whom liquidation and closure are practically impossible (e.g., "too important to fail") should not be granted the complete autonomy of borrowing. Any borrowing of a PSU should be reviewed and approved by the Ministry of Finance or local Bureau of Finance. 3.24 Deepening the reform of personnel and compensation systems. Third, further reform of personnel and compensation systems holds the key to improvement in incentives. In light of service standards and resource agreements mentioned above, each PSU should review its staffing level and consider options for retrenchment and reassignment, including early retirement and voluntary separation. A realistic reform plan needs to be designed to clean up all informal salaries, bonuses and other compensations by eliminating fraudulent and illegal incomes and regularizing others. 3.25 Empowering clients. Fourth, better accountability and performance management could also be achieved by empowering clients. Involving the public in service delivery monitoring and evaluation i s one way to empower them. While client evaluation has already taken place on a considerable scale, improving their methodology and quality, making these evaluations independent, and publishingthem would enhance the competition among service providers, and through that the quality and cost- effectiveness of services. Other countries have used "citizens' charters," "scorecards," "league tables'' and other tools to compare facilities in this way.28 It i s also worth considering introducing competition among providers by shifting financing from direct subsidization of providers (such as a school) to funding of the users (such as parents), and allowing them the choice of provider. The experiments in health insurance could include such features, and moving to a capitation grant specifically designed for children of targeted groups (e.g., migrant workers) in education could do the same. In terms of regulation, the government may consider reviewing the competitiveness and contestability of the market for services, sector by sector and region by region, to identify monopolistic powers, and implicit and explicit privileges authorized by the government. Taking into consideration the particular situation of the sector and the region, such 27 Two new divisions have beencreated within MOF in late 2004 to handle this issue. 28 A citizens' charter specifies the provider's promises in terms o f standards o f service delivery. A scorecard is a means to evaluate a service provider on objective performance criteria and/or client's perception of service quality. A league table ranks service providers such as hospitals based on their performance on scorecards. League tables o f local governments can also be constructed based on the performance of local governments across a range of services. Towardsa PSU Reform Strategy 45 monopolies and privileges should be either removed as soon as possible, or put under close and transparent public scrutiny to minimize their damage to the fairness of competition. ORGANIZEREFORMS 3.26 Setting up u tusk force. The central authorities need to give considerable thought to the organization, principles, planning, and phasing of the reforms. Given the sheer number of PSUs, and the numerous sectoral reforms already ongoing, reforms are unlikely to be implemented in a centrally managed way-one size does not fit all. Nevertheless, central guidance, coordination, and capacity for reforms are highly desirable. As a first step, the central government could set up a task force for PSU reforms directly under the State Council that would include core agencies such as NDRC, MOF, SCOPSR, MOP, Ministry of Civil Affairs, and major line ministries such as MOE and MOH. Inclusion of local government representatives can be considered as well. The task force should, among other things: i. CanyoutasurveyofPSUstodeepentheinformationavailableonPSUs. While the available MOF database contains lots of useful information, it still lacks the detail needed to define an overall strategy. It i s advisable therefore that the central government task force organize a comprehensive survey to fill the knowledge gap before any decision on major reform actions i s taken. The survey could be designed by the central government task force in consultation with relevant agencies and conducted at each level of the government. Major efforts could be devoted to ensure that PSUs at the county and township levels are well covered. .. 11. Design an overall strategy for reforms that includes principles for reform, phasing, timing, and responsibility for reforms. At the top of the government hierarchy, overall strategy and general guidelines of the reform must be endorsed by the top leadership. As some senior policy advisors have indicated, there may be a need for a formal "decision" to be adopted by the CPC Central Committee and the State Council on PSUreform. The central government task force can then be mandated to play a leading role in coordinating regional reform programs. They can be charged with the responsibility to ensure regional programs are in line with national strategy and guidelines and endorsed by key functional and sectoral ministries. Experimentation could be made compulsory in every province. Before major reform actions are taken province-wide, experiments could be carried out in a few selected counties to ensure that key social and economic risks are identified and properly provisioned for. Given the potential social impact of the reform, urgency must be tempered with caution to ensure reform measures at least do no harm. iii. Preparedraftlaws,regulationsandproposalsofchangesinlegalrules. Thiscould include a more detailed assignment of functions among levels of government to be included in the forthcoming revision of the budget law. Also included would be preparation o f legislation for new to-be-created types of PSUs, and amendments to laws and regulations, in particular government procurement rules, to allow NPOs to become providers of public services contracted by government. 46 Deepening Public Service Unit Reform to Improve Service Delivery iv. Prepare the framework of performance management. This would require a definition of a methodology for determining performance standards and design of monitoring and evaluation of those standards. It also involves drafting standard resource agreements for individual PSUs to be used by supervising ministries and local governments to define individual performance standards and use of budgetary and extra-budgetary resources. v. Coordinate reform actions of ministries and local government. This would involve a definition of the responsibilities of line ministries and local governments in implementing the reforms, and specify timing and phasing of reforms as well as the reporting requirements to the task force and the State Council. 3.27 Individual ministries and local governments are likely to benefit from setting up their own task forces to implement the reform for their region or sector, and the central task force could be a resource for these entities, and a depository of analysis and knowledge on PSUperformance. However, continued central guidance and oversight i s desirable, and probably needed to keep up the momentum of reform. 3.28 Adopting a more decentralized review and classification approach, The traditional classification approach adopted in past reforms i s not adequate to differentiate the role of the government across services of varying economic nature, nor is it capable of capturing the sectoral and regional diversities. To better reflect the complexity and diversity of PSUs in a reform strategy, the government i s advised to adopt a more decentralized review and classification approach, which essentially requires each level of the government to review all relevant services delivered by PSUs, specify the desirable role of the government in their delivery in terms o f regulation, financing and provision with a clear view on what kind of market failure i s present. All these should be done under general guidelines (including minimum service standards) set forth by the central government task force. MANAGEIMPACTONEMPLOYEES THE 3.29 Reform of PSUs i s likely to have significant impact on employees. I t i s probably inevitable that some reform measures that are necessary for improvement of public services may negatively affect the jobs, salaries, bonuses, and pension benefits of some PSU employees, at least in the short term. Managing the impact on employees i s therefore critical for the success of the reforms. 3.30 Communicating with employees on the limited scale of potential layofls. Given the experience of SOE reform over the past decade and the popular perception of overstaffing in the PSU sector, the launch of a major-scale reform in PSUs may easily raise concerns of a new round of mass layoffs. To guard against this risk, timely communication between the government task force on one hand and PSU employees and the general public on the other hand, is critical. Employees of PSUs should be kept fully informed about the development of PSU reform strategy and the implications for them. Implementing PSUreforms requires governments at all levels to understand the potential labor market impacts of the reform measures and to design social protection mechanisms to meet the needs. Detailed quantitative estimates of the impact cannot be made until Towardsa PSU Reform Strategy 47 needed data are collected in the survey recommended above, and the classification process i s completed. Nonetheless, based on the nature of services provided by PSUs and the characteristics of its labor force, an approach that i s different from SOE reform can still be argued for. It i s conceivable that the government i s not facing the pressure of mass layoffs at the aggregate level or devastating employment pressure in certain regions, as was the case in SOE reforms. 3.31 Adopting a gradual approach to rationalize PSU labor force. The process of shedding redundant PSUemployees could be more contained and more gradual to ensure social stability than in the SOE reform case. One option could be to follow the pilot initiatives to explicitly introduce employment contracts, with potentially redundant employees being treated in some differentiated ways. For a predetermined period of time, the redundant employees can be given basic protection in the form of basic wages. Inthe meantime, active measures need to be taken to provide information and help them to identify alternative job opportunities. After a certain pre-specified period, those individuals who still cannot find a job shall be given a severance package and be transferred to the unemployment insurance program. Meanwhile, one should hasten to add that before such an arrangement i s introduced, it i s essential to close off the current system. 3.32 Being both determinedand realistic in strengtheningfinancial accountability. Strengthening financial accountability and deepening the reform o f personnel and compensation systems are all crucial in getting the incentives o f PSUs right. However, some PSU employees stand to lose if reform i s carried out in the way that is required by improvement of incentives. Past reform experience suggests a high risk of the reform beingjeopardized by strong resistance from potential losers. One particular aspect of this issue, namely regularizing extra-budgetary and off-budgetary financing, deserves special attention. Such financing, as suggested in Chapter 1, i s often the source o f funds which allows for an increase in compensation of PSU employees. In many cases, the benefit i s extended to officials of their supervisory departments such as line bureaus or ministries. It is defacto part of PSUs' financing, as well as the supervising department's financing. Simply ignoring this in the reforms may be unrealistic, given that it perhaps accounts for a fairly large share in total compensation of PSU employees and some officials of their supervisory departments. At the same time, providing budget financing to compensate for the possible losses of such financing may not be affordable, or desirable if the bonuses paid with this money are out of line with the official salary system. The PSU reform task force and the government face the challenge of designing a realistic approach that minimizes resistance to reform from potential losers, and in the meantime stays firmly committed to implementation of the reform. A comprehensive review of pay scales inthe public sector could provide a way forward. 3.33 Aiming at a unified pension system. PSUpension reformi s an urgent issue for those PSUs that are slated to be converted into enterprises, and a long-term challenge that has implications for improvement of public service delivery through its impact on the functioning of the labor market for PSU staff. Inthe long run, the fundamental questions that the government needs to answer are: "Should establishing a unified reformed public pension system be the strategy objective? If so, what should be the key feature of the system and the path forward?" The right answer to the first question lies in a consistent 48 Deevening Public Service Unit Reform to Improve Service Delivery national pension policy framework that applies to the entirety of the Chinese labor force in the long term, including civil servants andPSUstaff. Duringa long transition period, dual systems and parallel reforms will needto be in operation simultaneously. It i s worth highlightingthat embarking on agradual unificationprocessrequires further rationalizing of the public pension system for enterprise employees, carrying out parallel reforms to ensure adequate return to pension investment, and developing a financing strategy to support the transition. REFERENCES Cheng, Siwei (edited): China's PSU Reform: Choice of Models and Guidelines based on Classifications, Democracy and Construction Press, Beijing, 2000. Cheng, Siwei: Studieson Economic Reforms and Development in China,Oxford University Press, 2001. Fan, Hengshan: "Thoughts on Some Important Issues Concerning China's PSU Reform," speech made at the International Seminar on InternationalExpedence of Public Service Reform and China's PSUReform, Beijing, March 23, 2004. Ge Yanfeng: "Some Principle Issues to be Studied and Resolved inPublic Service Unit Reform," Management World,No. 1,2003. Gill, Derek: "Signposting the Zoo," OECD Budgeting Joumal, Vol. 2, No. 1,2002. Hansmann, Henry B.: "The Role of Nonprofit Enterprise," YaleLaw Joumal, Vol. 89 (April) 835-898, 1980. Li,Ling: "Public ServiceUnitReforminChina: HealthSector," backgroundpaper preparedfor the World Bank, 2004. Ke, Bingsheng: "Study on Agriculture Extension Services inChina," background paper preparedfor the World Bank, 2004. Koppel, A., K.Meiesaar, et al: "Evaluation o f primary health care reform inEstonia," SOCSci Med 56 (12): 2461-6, 2003. M O E (Ministry of Education): Annual Report of National Education Finance Development, MOEFinanceDepartment and Shanghai Intelligence Institute, 1997. M O E (Ministry of Education): ChinaEducational Finance Statistical Yearbook 2003, Finance Department of MOE and Social Department of NBS, China Statistics Press, Beijing, 2003. MOF (Ministryof Finance): FinancialNorms of PSUs, 1996. Available at: http://www2.scut.edu.cn/audit/kjifg/caizbu-siyedanwei-caiwuguize.htm. MOH(Ministryof Health): National Health Accounts, 2002. MOST (National Research Center for Science and Technology for Development, Ministry of Science andTechnology): Reform of China'sPublic Institutions: Retrospect and Prospects, October 2003. NBS: China Statistical Yearbook,Beijing: China Statistics Press, various years. 49 50 Deepening Public Service Unit Reform to Improve Service Delivery NBS. 2003b: China SecondNational Census of Basic Units. Beijing: China Statistics Press. NDRC (National Development andReform Commission) PSUTeam: "A Statistical Study on Public Service Units," internal report presented ina workshop, June 2004. OECD: Distributed Public Governance: Agencies, Authorities and Other Government Bodies, 2002. Pilichowski, Elsa: "The reform of Public Service Units in China-the experience of OECD member countries in organizational change," background report prepared for the World Bank, March 2004. Polluste, Kaja. Ruth Kalda, Margus Lember: "Evaluation of Primary Health Care Reform in Estonia from Patients' Perspective: Acceptability and Satisfaction," Croatian Medical Journal, 45 (5): 582-587, 2004. Rose-Ackerman, Susan: "Altruism, Nonprofits and Economic Theory," Journal of Economic Literature, XXXIV, June 1996. Schick, Allen: "Agencies in Search of Principles, " Distributed Public Governance: Agencies, Authorities and Other Government Bodies, OECD, Paris, 2002. A Chinese language edition of this book i s also available. Schick, Allen: "Modernizing the State: Restructuring China's PSUs and Delivering Public Services," backgroundreport prepared for the World Bank, August 2004. State Council. 1998a: Tentative Regulation on Public Service Units Registration, Decree 252, Oct. 25, 1998. State Council. 1998b: Tentative Regulation on Registration and Administration of Min- ban Non-Enterprise Units, Decree 251, Oct. 25, 1998. State Council. 2000: State Council General Office Notice to Issue the Instructive Opinions on Urban Medical and Health System Reformby SCORES and Other Departments, Feb. 21, 2000. http://www.ceilaw.com.cn/dcdOOl/owa/cei.fgdetail-query?incode= 112708200002 Sun Xiaobing: A Study on Organizations and Institutions of China's Min-ban Education, China Youth Press, Beijing, 2003. Vito Tanzi and Ludger Schuknecht: Public Spending in the 2dhCentury, Cambridge University Press, Cambridge and New York, 2000. Wang, Rong: "A Study on Reform of Public Service Units inChina's Education Sector," background paper prepared for the World Bank, May 2004. References 51 World Bank: China 2020: Development Challenges in the New Century, Washington, D.C.. 1997. World Bank: China: Managing Public Expenditures for Better Results, Washington, D.C., 2000a. World Bank: Making Transition Workfor Everyone: Poverty and Znequality in Europe and Central Asia, Washington, D.C, 2000b. World Bank: China: National Development and Sub-National Finance, Washington, D.C., 2002a. World Bank: Transition: TheFirst Ten Years.Washington,D.C., 2002b. World Bank: World Development Report: Making Services Work for Poor People, OxfordUniversity Press, New York, 2003. World Bank: Citizens, Politicians and Providers: The Latin American Experience with Service Delivery Reform, draftreport, 2004.