Lessons from International Experience on Fiscal Decentralization for Regional Governments Volume I: Main Report prepared under the World Bank Partnership RAS for Chile1 1 This report was written by Jorge Martinez-Vazquez (Regents Professor, Georgia State University) and Ewa Korczyc (Senior Economist, World Bank). The work benefitted from comments by Pedro L. Rodriguez (Lead Economist, World Bank), Fernando Blanco (Principal Economist, IFC), Ayah Mahgoub (Senior Urban Development Specialist, World Bank), and the Chile DIPRES team led by Álvaro Miranda and Rodrigo Montero. Contents Foreword ....................................................................................................................................................... 4 Executive Summary ...................................................................................................................................... 5 1. Chile country context .......................................................................................................................... 12 2. Motivation: What is the purpose of decentralization and does it work? – A brief review of decentralization outcomes ........................................................................................................................... 12 3. Fiscal decentralization: A Framework for the Assessment of Fiscal Decentralization System .......... 19 4. Regional governments position in the decentralized government structure ....................................... 21 4.1. Two vs. three-tier subnational government structure .................................................................. 21 4.2. Hierarchical dependency of subnational governments ............................................................... 23 4.3. Types of decentralization and de-concentration versus decentralization in parallel structures .. 24 5. Fiscal discretion .................................................................................................................................. 27 5.1. Assignment of expenditure responsibilities ................................................................................ 27 5.1.1. General principles of expenditure assignments and international experience .................... 27 5.1.2. International experience in expenditure assignments.......................................................... 29 5.1.3. Addressing the assignment of capital investment responsibilities ...................................... 31 5.1.4. How to go about assigning responsibilities to regional governments? ............................... 32 5.2. Generation of income and revenue autonomy ............................................................................ 35 5.2.1. Own revenue sources .......................................................................................................... 36 5.2.2. Transfers and grants ............................................................................................................ 46 5.3. Subnational borrowing and fiscal rules ....................................................................................... 59 5.3.1. Subnational borrowing ........................................................................................................ 60 5.3.2. Fiscal rules .......................................................................................................................... 63 6. Fiscal accountability and transparency ............................................................................................... 69 6.1. Evolving role of central government in decentralized countries................................................. 69 6.2. Budget process for regional governments ................................................................................... 70 6.3. Audit, control and evaluations .................................................................................................... 71 6.3.1. Audit and control ................................................................................................................ 71 6.3.2. Evaluation function ............................................................................................................. 72 6.4. Transparency ............................................................................................................................... 75 6.5. Citizen participation .................................................................................................................... 76 6.6. Decentralization as means to address regional grievances.......................................................... 76 7. Implementing decentralization reform ................................................................................................ 77 8. Lessons learnt on fiscal decentralization............................................................................................. 77 Bibliography ............................................................................................................................................... 81 Annexes ...................................................................................................................................................... 85 Annex 1. Practical steps in expenditure assignments.............................................................................. 85 Annex 2. Main responsibilities of regional/state government level in Latin America ........................... 89 Annex 3. Salient issues in the political economy of decentralization ..................................................... 92 List of Boxes Box 1. Why are countries undertaking regional reforms? .......................................................................... 22 Box 2. Most Common Problems with Expenditure Assignments............................................................... 29 Box 3. Representative Assignments of Expenditure Responsibilities ........................................................ 30 Box 4. Breakdown of responsibilities at the INTERMEDIARY and REGIONAL level ........................... 30 Box 5. Revenue autonomy .......................................................................................................................... 35 Box 5. Criteria for selecting tax revenues at the regional level .................................................................. 40 Box 6. Vertical fiscal imbalances ............................................................................................................... 47 Box 7. Horizontal fiscal imbalances ........................................................................................................... 47 Box 8. Adjusted population index............................................................................................................... 50 Box 9. International experience with streamlining of transfer system........................................................ 52 Box 10. Specific Earmarked Transfers Can Lead to significant Distortions .............................................. 52 Box 11. Types and effects of conditionalities ............................................................................................. 54 Box 12. International Practices and Examples in the Design of Capital Transfers .................................... 57 Box 13. Municipal Development Funds and Intermediaries* .................................................................... 62 Box 14. Fiscal Councils .............................................................................................................................. 64 Box 15. Rainy Day Funds ........................................................................................................................... 65 Box 16. Circumventing borrowing rules..................................................................................................... 66 Box 17. Link between National and Subnational Budgets ......................................................................... 71 Box 17. Experiences in institutionalizing an evaluation function for public policies................................. 74 3 Foreword The World Bank has been requested to accompany the Chilean government in conducting a comprehensive analysis of international experience with fiscal decentralization processes for regional level of subnational governments. Such analysis will provide lessons and recommendations that can assist the Ministry of Finance in designing and implementing the fiscal decentralization framework for regional governments in Chile. This work is aligned with the Government’s decentralization strategy which places the re gionalization process, in particular democratically elected self-governing regions with resources and responsibilities, at the center of the reform effort. This work is very timely, as Chile - confronted with social unrest towards the end of 2019 - will hold a referendum regarding the current constitution and is embarking on a national debate to redefine the fundamental principles of state organization. The study is comprised of two volumes. Volume I provides a systematic review of the relevant international experience in fiscal decentralization for the mid-level of government. The work focuses primarily on the aspects of fiscal decentralization from the fiscal discretion angle, specifically expenditure responsibilities, revenue assignments, transfers and subnational borrowing. This work is supplemented with a brief overview of selected aspects of fiscal accountability related to the budget process, audit and control; as well as transparency and citizen engagement. Volume II includes in-depth case studies with descriptions and analyses of decentralization processes in four selected countries: Colombia, Indonesia, Poland and Spain. This volume is prefaced by an overview that summarizes the main issues and draws relevant lessons from the experience of these four countries. Executive Summary Over the last several decades many countries around the globe have decentralized devolving fiscal, administrative, and political powers to sub-national governments at the regional and local levels. This trend toward decentralization has been motivated for a variety of reasons, including bringing government closer to the people and invigorating democratic institutions, a more efficient and leaner public sector, disenchantment with planning and centralized policies, and allowing for more autonomy to dissuade centrifugal forces and regional conflicts. Building on the economic rational, the basic premise of fiscal decentralization is that if decision-making and service delivery are moved closer to the population, this can increase the quality and quantity of public services provided for the same amount of resources spent. Investigating the impact of fiscal decentralization is a complicated task. There are several technical reasons why in some cases there is less than complete consensus on the impact of fiscal decentralization. One is that the level of fiscal decentralization existing in a country is not a straightforward issue to measure. Fiscal decentralization is multi-dimensional, involves expenditure and revenue assignments, effective autonomy for spending resources and collecting taxes, and so on. A second issue is the potential endogeneity or reverse causality; for example, fiscal decentralization may affect economic growth in the country but fiscal decentralization itself may be affected by how the economy is growing. Not addressing the reverse causality can distort the estimates of impact. A third issue is the lack of uniform data across countries, capturing the multidimensionality of decentralization, which can also lead to inaccurate estimates of impact. Bearing in mind the methodological challenges, overall, the international experience shows that fiscal decentralization can generate many positive outcomes. There are numerous studies that show a positive link between fiscal decentralization and service delivery. As part of the existing evidence, a growing number of household surveys typically have shown increased citizen satisfaction with decentralized delivery of public services. The literature on the effect of decentralization on infrastructure spending, and its composition and quality continue to develop. Recent studies as well as the earlier findings suggest that infrastructure in decentralized settings is of better quality and completed at lower costs than in centralized ones, but with some country exceptions, especially in the LAC region for example Peru and Mexico. The literature also provides substantial evidence of positive impact of fiscal decentralization on issues like macroeconomic stability, inequality, poverty, and regional disparities. Nonetheless, potential pitfalls of fiscal decentralization need to be emphasized as well. A multilevel provision of public services may result in inefficient expenditure levels. This particular risk may take the form of under provision, or that decentralizations ends up with inefficiently low levels of services, if for example, in a “race to the bottom,” subnational governments compete among themselves to attract businesses etc. by continuously lowering taxes and end up with much smaller budgets. The inefficiency in service levels may also take the form of too much expenditures, as when subnational governments compete among themselves by spending more and more on infrastructure and other means to attract businesses, or when they find it politically easy to spend every penny they get in transfers (which is known in the specialized literature as the “flypaper effect of transfers” since money sticks where it hits). Among other potential pitfalls, fiscal decentralization can also be associated with several manifestations of the “common pool” problem which is related to the presence of soft budget constraints and moral hazard which in turn lead to national bail-outs due to excessive subnational borrowing. This report uses a conceptual framework developed by the World Bank staff for the assessment of fiscal decentralization arrangements across client countries, in which fiscal discretion and fiscal accountability are the main building blocks. Fiscal discretion can be divided into four elements, namely, the assignment of expenditure responsibilities, revenue generation, intergovernmental transfer systems, and local government borrowing. The framework emphasizes that an authority can be held accountable for performing a specific function only if it has the fiscal resources and the discretion to perform that function. Similarly, the framework argues that accountability is not an automatic outcome of increased discretion and that governments need to make a conscious effort to create structures that would hold local governments 5 accountable. The two essential elements of accountability include public accountability, where responsible individuals are held accountable by other elected or non-elected officials, and social accountability, where public officials are answerable directly to the citizens. Fiscal discretion aspects Expenditure responsibilities Internationally, expenditure assignments for subnational governments are usually guided by specific principles including: correspondence, subsidiary and benefit principles. It is important to note that the application of the principles above facilitates the assignment of expenditure responsibilities to different levels of government but does not always yield an unequivocal answer. The international experience provides some valuable lessons on what to avoid as poor practice in expenditure assignments. First, the most commonly found problem is the lack of a formal explicit assignment of responsibilities. A second common problem in the assignment of expenditure responsibilities is the presence of the wrong or inefficient assignments, which openly brake with the principles above. A third common problem is the extensive use of concurrent responsibilities (those assigned simultaneously to two or more level of government) without clarifying the level of government responsible for the different “attributes” of the service: regulation, financing and implementation of those services. This is often the most common cause of confusion, friction and inefficiencies. The international experience (including developed and developing countries, with constitutional federal and unitary regimes) shows great diversity in expenditure assignments, which reflects the different approaches to decentralization and other relevant factors, including the federal versus unitary nature of the state, history, geography or level of development. A fair conclusion is that “there is no single best assignment.” Also, what may be considered a good assignment is likely to change over time with changes in the costs of provision as well as technological innovations and changes in citizens’ preferences. Nevertheless, there is a need at any given moment in time to have a concrete, clear and stable assignment of expenditure responsibilities. Revenue autonomy and transfers Revenue autonomy is at the core of fiscal decentralization. The fundamental premise of fiscal decentralization is that it increases the overall efficiency of the public sector and leads to improved social and economic welfare. One critical assumption behind those results is that decentralized governments will generally be more accountable and responsive to citizens’ needs and preferences. At the same time, there is a general agreement among experts in decentralization that the increased accountability associated with decentralization can only be assured when subnational governments have an adequate level of autonomy and discretion in raising their own revenues. The logic is simple but powerful: subnational authorities will spend resources differently when they need to incur the political costs of taxing their constituents, in comparison to what they would do if the money comes to them, practically cost free, from the central government. This is the argument behind the “tragedy of the commons.” When subnational expenditures are being financed by other taxpayers outside the jurisdiction, there will be perverse incentives to overspend and/or to spend poorly. There are multiple trade-offs in designing the tax structure at the subnational level. Given the different criteria that subnational taxes should meet, the choice of particular taxes comes down to a comparison of the advantages and disadvantages that each of those taxes offer. Among the desirable properties are: good revenue potential, with high elasticity or buoyancy (meaning that revenues will grow proportionally with the economy), high adaptability to the benefit principle (so that those paying the tax are those using the public services), an even geographical distribution of tax bases (so that there are less disparities in per capita revenues across jurisdictions), fairness and equity in the distribution of tax burdens (those with higher incomes pay proportionally more), lower costs of administration and compliance by taxpayers, higher acceptance and better perception about the tax among officials and the public, and high visibility of the 6 taxes so that can facilitates greater accountability by public officials. Among those properties that are negative, there is inter-jurisdictional mobility of the tax bases, negative efficiency effects in the allocation of economic resources, high sensitivity to fluctuations with the business cycle in the economy, the risk of incentivizing or facilitating corrupt practices among tax administrators and taxpayers, and the risk that tax burdens may be exported to other taxpayers in other jurisdictions (so bringing less accountability and fiscal responsibility). Apart from subnational taxes, intergovernmental transfers like revenue sharing and equalization transfers, play an important role in the resource envelope for subnational governments. Different types of transfers are necessary to help close vertical and horizontal imbalances stemming from either inadequate or unequal own resources (through revenue sharing and equalization transfers). Revenue sharing following devolution objectives clearly benefits the most economically dynamic regions of the country. But the larger inequality or horizontal fiscal imbalances can be addressed—as is generally the case in the international practice— with a separate instrument, i.e. a strong equalization grant. Well-designed equalization grants recognize automatically in their formula revenue sharing funds (as well as own revenue generation capacity) as part of the fiscal capacity calculation, which reduces, if not eliminates, the equalization funds that “richer” subnational governments potentially receive. Capital transfers to subnational entities are also part of the revenue fiscal architecture. International experience with the design of capital transfers shows a large variety of approaches that are used globally. Most countries use some form of capital transfers in support of subnational governments for specific sectoral expenditure areas such as roads, water and sewage treatment plants, transportation, housing, education, health, and so on. Country experiences with regards to the design features of capital transfers vary along several dimensions. Regarding the mechanism used to allocate capital transfers, country experiences vary from ad hoc allocation decisions to formalized approaches using pre-established formulae. Similarly, in terms of the degree of flexibility in the use of capital grants, country experiences vary from the least flexible “project-based grants” to unconstrained funds provided as part of general revenue transfer. Subnational borrowing and fiscal rules Subnational borrowing and fiscal rules can potentially play an important role in safeguarding macroeconomic stability and sustainability of subnational public finances. A number of factors create adverse incentives for subnational governments to overspend, undertax, and/or borrow excessively – all root causes of fiscal imbalances. These excesses can compound fiscal profligacy by the central government, or undermine efforts to ensure fiscal discipline, thereby jeopardizing macroeconomic stability and, possibly, debt sustainability. The most important risks are related to the common pool problem and moral hazard. Fiscal and subnational borrowing regulations rules can help prevent subnational fiscal autonomy from undermining sustainability at the macro level for example through excessive subnational borrowing. Where they are allowed access to capital markets, sub-central governments without significant revenue raising autonomy are more likely to be perceived by lenders as borrowers that are implicitly guaranteed by central government. The costs to the rest of the country of profligate fiscal policy can thus arise through the moral hazard of bail-outs from central government, rising overall borrowing costs and higher and more volatile public spending and tax rates. Both numerical and procedural rules can be helpful in fostering fiscal discipline at the subnational level but will not be effective if the basic architecture related intergovernmental fiscal relations are poorly designed and undermine incentives for fiscal responsibility. Fiscal accountability aspects Accountability is not an automatic outcome of increased discretion of the local governments. Specific mechanisms should be designed to ensure that citizens and higher officials are able and willing to hold local governments accountable for their discretion. The notion of fiscal accountability relates to issues related to public accountability, through well-established relations between the central government and subnational 7 governments, properly designed budgeting process, audit, control and evaluations, as well as social accountability through transparency and citizens’ participation. The process of decentralization involves a changing role of the central government. In the decentralized environment, the focus of the central government is increasingly on putting in place the strategies, framework conditions and regulations, incentives – rather than in the implementation of policies. The central government has also an important role to play in ensuring that decentralization does not widen disparities in terms of access and quality to the public services provided by subnational governments in their new functions. Finally, the central government also needs to renew its supervisory and monitoring role vis-à-vis subnational governments, especially in legal and fiscal matters. Effective decentralization strategies require that the subnational government strengthen its internal control and audit instruments while devolving more power to local governments. An effective, efficient, transparent, and rules-based PFM system is an essential tool for a government in the implementation of a fiscal decentralization program. Within the PFM framework, internal control and audit functions support fiscal decentralization process by promoting transparency and accountability in the use of public resources, ensuring allocation of public resources in accordance with citizens’ priorities, and supporting aggregate fiscal discipline. In addition, given the increased responsibility of subnational governments to deal with resource allocation and service provision, the importance of transparency at the subnational level has increased dramatically. Transparency at the subnational level is required in order to expose service delivery issues, which are a major concern in many countries. Moreover, subnational budget documents deal with items and expenditures that are particularly visible and tangible for citizens. In this sense, subnational transparency has the potential to provide answers to basic questions such as what happened to central government funds at the local level and whether plans to build a new neighborhood school or carry out local road works were actually implemented. Also, social accountability mechanism for decentralized governments can be strengthened through platforms allowing for citizen participation. There are different models how this can be achieved. Some of the common examples include citizen monitoring committees and participatory budgeting. Implementation Aspects Undertaking comprehensive fiscal decentralization reform faces lawmakers with the choice of either sequencing those reforms over time or clustering the reforms in a short period of time by simultaneously adopting all, or the most important, pieces at once. A sequenced gradual implementation of reform is generally thought to be more desirable than a “big bang” approach because it carries substantially lower risks in terms of the information available at any time and the ability to spread the costs of the reform over time. A sequenced approach may also allow for some experimentation and being able to adjust to different initial capacities of subnational entities. However, gradualism and a paced implementation should never be interpreted as not requiring a comprehensive blue print for decentralization reform. Proper sequencing demands a roadmap with explicit plans and goals, accompanied by the institutional and administrative changes necessary to accomplish those plans and goals. The introduction of important reforms also requires adopting transition measures for their differential impact on subnational entities. Most fiscal decentralization reforms imply that there will be winners and losers. Politically or relatively economically powerful subnational governments may threaten the reform unless their concerns are addressed. Here, policy makers can select to go with a “hold harmless” approach, where no entity end up with less resources it had before the reform, or with a so-called “cold turkey” approach, where the new system is adopted no matter what, or, finally, with an intermediate approach, where the reform are introduced over a period of time and losers are at least partially compensated. The 8 political economy of reform in most countries often calls for some form of sequencing with compensation, if not a strict “hold harmless” approach. One last dimension of any successful decentralization reform effort is the planning and implementation for monitoring and evaluation, which should lead to periodic fine-tuning of institutions and processes, as well as to the provision of technical assistance to subnational governments in those areas where they are underperforming. Carrying out successful monitoring and evaluation typically requires a unit, most likely located in the Ministry of Finance, with trained staff exclusively dedicated to this task, and an extensive information and data system that will enable proper monitoring and evaluation. International good practice and lessons for Chile There are several basic principles that should guide decentralization reforms in general to ensure accountability, transparency, and an efficient performance at all levels of government. Those principles are listed below, and an attempt has been made to relate them to the Chilean case; however, because the Chilean process is still fairly new and not quite defined, the comments will necessarily remain general: • Fiscal decentralization policy should be comprehensive. Ideally, the design of a decentralization program should address all of the components of the system even if not all components are implemented right away. Successful decentralization policies build on more than the traditional fiscal issues, including an electoral system and other forms of political accountability and a capable civil service. Piecemeal reform, encompassing only one element of the system, such as increased financing of local governments, is unlikely to deliver the benefits of decentralization. For Chile this means that the creation of the new regional governments in Chile will demand a detailed “white paper” and road map covering all dimensions of decentralization: fiscal, political and administrative decentralization, and within the fiscal decentralization dimension including the architecture of the assignment of functional expenditure responsibilities, the assignment of own revenue sources, the system of intergovernmental transfers, and last, the prudential access to borrowing for capital investment purposes. The temptation will be just to do part of the work at this time and postpone other aspects to a later day. However, that will likely imply uncoordinated, if not dysfunctional, decentralization, creating problems that will be hard to solve down the line, as the process becomes increasingly path-dependent. Nevertheless, the requirement that the design should be comprehensive does not necessarily mean that the full architecture needs to be implemented all at once. This latter is not the case. However, there should be a clear roadmap for when and how the different elements of the reform will be implemented, accompanied by a full political commitment to do so. • Finance should follow function, which means that the first step should be the assignment of expenditure responsibility to subnational governments, and only following that the assignment of revenue-raising powers and central government transfers should be determined. Known expenditure responsibilities facilitate the determination of the level of funding and the types of revenue instruments: what services should be financed by user charges, local taxes, conditional grants, and so on. On the implementation side, the decentralization of expenditure competencies needs to be carried out in the context of reformed national sectoral laws that support and do not hinder the decentralization reforms. Subnational governments must have the capacity to deliver the assigned services or else the expenditure assignments may need to be asymmetric and restricted to those that have these capacities. Notice that determining the required level of funding demands a determination of the cost of providing these services. This can be done by adopting a “retrospective” method of using some form of per capita (or per client) financial norms or physical service standards. For Chile this means that no decision should be made about the financing sources for the new regions until there is clarity on what functional responsibilities or elements of those responsibilities will be assigned to the regions and what the expenditure needs associated with those responsibilities will amount to. The exercise of assigning expenditure responsibilities to the new regions will be demanding because some of 9 those responsibilities will be detached from the central government portfolio and some from the local governments’ portfolio. There will be also a need to review and update sectoral laws (on education, health, transport, etc.) in Chile as many of those laws and related regulations are likely to refer to the functional responsibilities of the new regional governments. • There should be clarity of roles between levels of government. Clarity, transparency, and stability are paramount for achieving the accountability at the administrative as well as the political levels that efficient governance requires. Much too often, overlapping responsibilities accentuated by expenditure dumping or unfunded mandates combine with frequent legislative and regulatory changes to create a diffused and confusing system of expenditure responsibilities that compromise efficient service provision. For Chile, from this perspective, the new legislation on regions should strive to bringing clarity to the assignment of functional responsibilities for the very start in an uncompromising way. These assignments preferably will be permanent or at least not often changed. • A significant measure of autonomy for subnational governments on both the expenditure and revenue side is crucial for realizing the efficiency potential of decentralized government. On the expenditure side, this would require providing subnational governments with budget flexibility to decide —within limits— expenditure priorities, the service levels offered in accordance with local preferences, and the best means to deliver those services. On the revenue side, a minimum of subnational autonomy to set tax rates on at least one significant tax source assigned to the subnational level; allowing subnational governments the powers to raise revenue is a requirement for accountability. Revenue autonomy also facilitates addressing vertical fiscal imbalances and building up credit-worthiness at the local level. Generally, subnational governments should not be allowed the powers to define tax bases. The best vehicle for local autonomy is discretion in setting tax rates for some meaningful taxes. While autonomy should be explicit and well- defined, it must also be circumscribed in some important respects, first of all with respect to the access to borrowing by subnational governments. From a Chilean perspective, the fundamental question to be asked is why there is a need to further decentralize government in the country by creating new regional governments. The answers are likely to include the need to bring government closer to the people, increase the efficiency of public expenditures, to further democratize the country and to achieve more balance growth and distribution of resources across the different regions of the country. However, decentralization will not bring the country much closer to any of those objectives unless the decentralization process provides significant measures of autonomy and discretion to the new regional governments and not only on the expenditure side, but also on the revenue side of the budget. • Decentralization should go hand in hand with an intergovernmental transfer system that has elements of equalization. Sustainable autonomy and economic efficiency, however, also require some equalization of opportunity so that local councils have an approximately “level playing field” with which to offer services to their residents. This observation points to the crucial importance of equalization transfers in fiscal decentralization design. In the design of intergovernmental transfers, the right order of policy formulation is to first ask and answer the question about which of many possible objectives, besides equalization, the intergovernmental transfer system is to accomplish. The major issue therefore is that different types of transfers have different objectives, and it is important to match grant design according to these objectives. From a Chilean perspective, the two most important lessons to be applied here are: first, to use a separate transfer instrument for each objective pursued by the central government so that its effectiveness can be monitored, and second, that equalizing expenditure needs and the fiscal capacity to provide for those needs across the regions in the country will not only be an equitable outcome that citizens demand, but it will be also a way to reduce inefficient migration in the country to its capital Santiago, simply because people are just looking for better access to basic public services. 10 • Subnational governments must face a hard budget constraint. If subnational governments are not forced to operate with balanced budgets, they become accustomed to looking to the higher-level governments to cover their shortfalls, they will consciously overspend, and will have little incentive to raise their own revenues. As many countries have learned, it is difficult to break the culture of a soft budget constraint. From a Chilean perspective, given the tight borrowing environment that has existed for local governments in the past, the most important lesson to derive here is that not allowing borrowing or having too little borrowing by the new regional governments can be as inefficient as having too much borrowing and soft budget constraints. The right, somewhat delicate, balance must be found so that the new regions can prudently invest in the public infrastructure necessary for their economic development, and in a manner that fairly distributes the financing costs for that infrastructure among the different generations of resident beneficiaries. • Institution building is a necessary companion of decentralization. A prerequisite for successful decentralization is that local governments possess the administrative infrastructure required to effectively carry out the responsibilities assigned to them, regarding the number and quality of staff, and equipment. Administrative capacity is part of the necessary institutions, but there are others that are equally, or more, important. Democratic representation, sound budget processes and institutional arrangements to ensure coordination and cooperation between different levels of government—both at the political and the technical level—are crucial for a multi-tier system of government to work efficiently. This reflects the fact that even if the structure of intergovernmental fiscal relations is very well designed, there will be a need to address the technical, administrative, and political issues that invariably arise in any system. From a Chilean perspective, it seems that the requisite political democratic institutions would be guaranteed. However, it is not so clear that the administrative capacity will be in place without a coordinated effort for training administrative personnel in their new roles and capacities. This element can make or break the success of the new regional decentralization in Chile. 11 1. Chile country context 1. In February 2018, Law 21073 was enacted to regulate the election of regional governors in Chile, fulfilling an important milestone in the history of the country. Upcoming regional elections are scheduled for 2020, and during the next 12 months, the work will be done on defining the conditions under which the new elected regional governors will operate. The current administration is preparing a draft Law that seeks to put in place financing conditions for regional governments and determine the level of fiscal autonomy. The Law is seen as a critical element of the fiscal architecture of regional governments and a starting point before the model of fiscal decentralization is fully developed. 2. The first democratic elections of regional governors will take place in the context of a broader national debate on the fundamental principles of the organization of the state in Chile. The widespread and long-lasting social unrest in late 2019 led to an agreement between the Government and the protesters to hold a referendum on the future of the current constitution and on the steps leading to drafting of a new one. It is very likely that the agenda of decentralization will feature prominently in future debates. As discussed in this report, decentralization reforms have been successful in many other countries in promoting democracy, representation and a sense of national unity and belonging, while, in turn, democratic regimes also tend to maintain higher levels of decentralization than authoritarian ones. By bringing government closer to citizens, decentralized systems promote, among other beneficial outcomes, regional convergence, and higher quality and more equitable access to public services. All these factors would seem to be very relevant for Chile as it embarks on elaborating a new model of state organization for the country. 2. Motivation: What is the purpose of decentralization and does it work? – A brief review of decentralization outcomes 3. Over the last several decades many countries around the globe have decentralized devolving fiscal, administrative, and political powers to sub-national governments at the regional and local levels.2 A frequently used index of regional authority reveals that more than two-thirds of the countries analyzed have carried out significant fiscal decentralization reforms (Hooghe et al. 2016). The Regional Authority Index (RAI) presents a useful way to explore trends in decentralization across many countries. This index is a comprehensive tool to measure the real degree of power of intermediate governments, going beyond fiscal indicators. The RAI specifically focuses on regional government, which is defined as an intermediate tier of government between the lowest, local tier and national government, with at least 150,000 inhabitants per regional unit on average. The RAI’s ten dimensions include fiscal autonomy and borrowing autonomy, but also lawmaking and executive control. RAI-scores are available for American, Asian and European countries since 1950. A striking observation is a clear trend of increasing regional authority across the three groups of countries and this trend is especially noticeable from 1970 onwards. Average regional authority was 55% higher in 2010 than in 1950. Of the 81 countries covered by the RAI, 52 experienced a net increase in the degree of regional authority and only nine experienced a net decline 2This section draws on Martinez-Vazquez et al. (2016). Throughout this section references are made to the evidence in numerous papers regarding the impact of fiscal decentralization. For space reasons the specific references to each of those papers are omitted in this report but they can be found in the reviews contained in Martinez-Vazquez et al. (2016). Figure 1. Regionalization in America, Asia and Europe since 1950 Source: OECD (2019) 4. The trend toward decentralization has been motivated for a variety of reasons, including bringing government closer to the people and invigorating democratic institutions, a more efficient and leaner public sector, disenchantment with planning and centralized policies, and allowing for more autonomy to dissuade centrifugal forces and regional conflicts. According to the OECD3, the rising global decentralization trends can be attributed to a combination of three main categories of factors: political drivers, economic drivers and megatrends (Figure 2). • Political motives: decentralization reforms are often pursued as part of the broader political reform process, for example related to territorial development reforms, reforms of the judiciary, civil service and regulatory frameworks, while also building greater accountability and reducing political instability. Decentralization is enshrined in several national constitutions, as a fundamental principle to deepen democracy, improve the balance of power between higher and lower levels of government and ensure the effective participation of citizens in decision-making for example in Austria, Belgium, Colombia, Germany, Mexico, Peru, Poland and Switzerland. • Economic drivers: are usually associated with the desire to increase efficiency in public service provision, to better use public resources and to ensure spending effectiveness (particularly in the context of public finance crises and against a backdrop of tight budget constraints), and the opportunity to increase equity in the access to public services and tailor policies to local contexts and population needs. A decentralized approach is also promoted in the context of regional development that is bottom-up driven and related to regional circumstances. For example, the economic drivers have played an important role in devolving the health care function to subnational level in India. Economic drivers also supported the recent decentralization reforms in Greece (the Kallikratis reform), where the goal was to streamline territorial organization and to reduce and optimize public spending by reinforcing decentralized service delivery. • Megatrends: the process of decentralization is also believed to be influenced by the global megatrends like ongoing urbanization, demographic changes, digitalization and globalization. 3 OECD (2019) 13 Regarding the latter, globalization has provoked a revival of local cultural identities, resulting in a growing economic and political power of subnational entities. In recent years, the role of nation- states has been re-emphasized in a context of backlash against globalization. In this context of democratic crisis in many countries, the role of local governments has also been emphasized as the closest level of government to citizens, and a way to better echo citizens’ demands and needs. Going “local” may be an answer to the forces that are “global” (Sharma, 2018). The megatrends have affected the process of decentralization in multiple countries. For example, the ongoing globalization and urbanization prompted Chinese authorities to devolve powers to local governments (cities) to attract new investment opportunities, foster local innovation and growth. Similarly, rising urbanization in Japan and Korea has been an important driver of deeper decentralization reforms. Figure 2. Drivers of decentralization Source: OECD (2019) 5. Building on the economic rational, the basic premise of fiscal decentralization is that if decision- making and service delivery are moved closer to the population, this can increase the quality and quantity of public services provided for the same amount of resources spent. To a large extent, these gains in efficiency are derived from the better information and greater accountability local officials have– as opposed to central government officials—about local needs and preferences.4 This change in governance systems, in turn, may affect resource allocation, economic growth and, ultimately, the overall welfare of the population. However, the specific nature of the linkages between decentralization and macroeconomic performance are both controversial and quite complex. This complexity in part follows from the many different forms that decentralization may take. 6. Of course, that decentralization may have benefits does not mean that most or all public goods and services should be decentralized. Some public goods have such large benefit areas that they should be delivered by the central government. There are also economies of scale in production and large externalities in consumption that justify the production and delivery of other public goods at the regional level. 7. In addition, not everything in decentralization policy, as the international experience well shows, is positive and desirable. For example, the cases of Argentina and Brazil decades ago show that poorly designed decentralized systems, for example the absence of a hard budget constraint for subnational 4 The efficiency benefits of decentralization depend to a large extent on citizens making their preferences known to decision makers, through “voice” or elections (Hirschman 1970), or by “exit” since taxpayers can vote with their feet by moving across jurisdi ctions (Tiebout, 1956). 14 governments, can lead to waste and macroeconomic instability. Also, basic institutional failure in issues such as representation and accountability or the presence of bureaucratic corruption can lead to the capture of government by local elites with overall perverse outcomes. Also, subnational governments may lack the requisite level of technical and administrative capacity to realize the potential gains from decentralization. 8. There are other potential pitfalls of fiscal decentralization that need to be emphasized. A multilevel provision of public services may result in inefficient expenditure levels. This particular risk may take the form of under provision, or that decentralizations ends up with inefficiently low levels of services, if for example, in a “race to the bottom,” subnational governments compete among themselves to attract businesses etc. by continuously lowering taxes and end up with much smaller budgets. But, the inefficiency in service levels may also take the form of too much expenditures, as when subnational governments compete among themselves by spending more and more on infrastructure and other means to attract businesses, or when they find it politically easy to spend every penny they get in transfers (which is known in the specialized literature as the “flypaper effect of transfers” since money sticks where it hits). 9. Among other potential pitfalls, fiscal decentralization can also be associated with several manifestations of the “common pool” problem. One such manifestation is the presence of tax collection laxity by subnational governments with high dependence on central government transfers. In such case, subnational authorities find it more convenient to be financed from (common) national resources than from the politically costlier collection of own revenues from their constituents. Another manifestation of the common pool problem is the presence of soft budget constraints, with national bail-outs due to excessive subnational borrowing. 10. Subnational governments presently are key actors in the delivery and provision of public goods and services to citizens, including social services such as education and health in many countries around the world. One key question in those countries that already have decentralized and those that are thinking of doing so is what is the impact of fiscal decentralization on the quality and quantity of service delivery? And beyond public services, what is the impact of decentralization on other economic and institutional outcomes such as, economic growth, macroeconomic stability, regional economic convergence, corruption, accountability and political representation, and so on. 11. In this section we briefly review what is known to date on the impact of fiscal decentralization along the above-mentioned dimensions and others. Because the literature on these issues is extensive, this brief review cannot do it justice and will necessarily focus on the main outcomes, while highlighting where there is consensus and dissent in the previous literature on observed outcomes. 12. There are several technical reasons why in some cases there is less than complete consensus on the impact of fiscal decentralization. One is that the level of fiscal decentralization existing in any particular country is not a straightforward issue to measure. Fiscal decentralization is multi-dimensional, involves expenditure and revenue assignments, effective autonomy for spending resources and collecting taxes, and so on. A second issue is the potential endogeneity or reverse causality; for example, fiscal decentralization may affect economic growth in the country but fiscal decentralization itself may be affected by how the economy is growing. Not addressing the reverse causality can distort the estimates of impact. A third issue is the lack of uniform data across countries, capturing the multidimensionality of decentralization, which can also lead to inaccurate estimates of impact. None of these issues will be further addressed here, but their mention serves as a cautionary note on the level of confidence that should be put on the findings we move to summarize below. 15 The impact on service delivery, infrastructure, and expenditure composition 13. Since the start of fiscal decentralization theory (Tiebout, 1956; Oates 1972), the fundamental claim has been that fiscally decentralized systems yield more efficient public service delivery, in the sense that government is more likely to deliver the services most wanted and needed by citizens. This higher so-called “allocative efficiency” is due to the better information decentralized governments have about those needs. A second decentralization concept is that of production efficiency; that is, delivering a bundle of public services at a minimum cost, then translating into an increased quality and quantity of the services. 14. Education and health are among the most important types of decentralized services and a lot of the empirical literature has focused on the impact on those two sectors, as reviewed in Martinez-Vazquez, Lago-Peñas and Sacchi (2017). Many researchers have found that decentralization affects education outcomes positively, whether the focus is on individual countries or cross-country studies. The findings for health services confirm that decentralization has similar virtuous effects, especially in middle- and higher- income countries. In some developing countries, however, decentralized governments may have given lower priority to healthcare services with respect to education services. There is also considerable consensus in the literature that countries that decentralize tend to have relatively higher spending on social sectors vis- à-vis other functions. Regarding other public services, the evidence on the impact of decentralization is a bit more mixed but leaning strongly towards positive impact on service delivery. As part of the existing evidence, a growing number of household surveys typically have shown increased citizen satisfaction with decentralized delivery of public services. 15. In the case of infrastructure there has been a great decentralization push worldwide (Frank and Martinez-Vazquez 2015). The share of sub-national governments in total capital expenditures of a country is typically twice their share in total recurrent expenditures, often reflecting the needs from urbanization. All in all, the decentralization of infrastructure spending brings significant intergovernmental coordination challenges. The literature on the effect of decentralization on infrastructure spending, and its composition and quality continue to develop. But the earlier findings that infrastructure in decentralized settings is of better quality and completed at lower costs than in centralized ones has been reconfirmed by more recent studies, but, nevertheless with some country exceptions, especially in the LAC region for example Peru and Mexico. The impact on economic growth, macroeconomic stability and fiscal sustainability 16. Because fiscal decentralization enhances economic efficiency, it may also have a corresponding effect on the dynamic setting of economic growth. Theoretically, more fiscal autonomy may be associated with higher output per unit of labor and therefore higher growth rates. However, a causation path is not clear-cut; for example, decentralization may affect growth indirectly through its impact on other socio- economic variables which are known to positively affect growth, such as macro stability and governance quality. From an empirical perspective, there has been hardly any other topic on the impact of decentralization that has received more attention in the literature. However, the findings remain mixed as reported. Some papers have found a negative or no impact. Many more papers, again based on country case studies and cross-county panel data, have found positive effects. It should be expected that studies using different data containing different institutional setups, estimation techniques and specifications and different definitions of decentralization may produce diverging results regarding the impact of fiscal decentralization on growth. But, clearly, the results in the empirical literature are not robust. 17. With regards to macroeconomic stability, there are good reasons to expect that poor decentralization design, for example containing soft budget constraints for subnational borrowing or large vertical imbalances with lithe own subnational revenues, may lead to challenges for a country’s fiscal 16 sustainability, as illustrated by the experiences of Argentina and Brazil with hyperinflation several decades back. On the other hand, decentralization itself could encourage governments’ fiscal discipline. Providing all levels of government with sufficient tax autonomy and fiscal rules would encourage fiscal responsibility. Although the empirical evidence on the relationship between decentralization and fiscal stability offers some mixed results, the overwhelming evidence is that well designed systems can enhance macro stability even in times of significant turmoil, such as with the recent worldwide events of the Great Recession (Lago et al. 2019). The impact on income inequality and poverty 18. Fiscal decentralization can directly affect poverty and income distribution among individuals (and not only across regions) by facilitating access to basic services, but also indirectly in many ways (by means of growth, expenditure composition, the quality of governance). Ultimately, those impacts will depend on the specific characteristics of each decentralization process. In addition, although closely related, income inequality and poverty do not need to move in the same direction in response to fiscal decentralization. A good number of studies find a beneficial effect of decentralization on the Gini coefficient, thus reducing inequality in the economy (Martinez-Vazquez et al. 2017). On the direct impact of decentralization on poverty, the empirical evidence is more mixed, with earlier finding positive outcomes and more recent ones finding negative effects, introducing doubts on championing decentralization as a poverty reduction process. The impact on geographical and interregional disparities 19. Fiscal decentralization may affect regional economic convergence and increase or dampen disparities in public services availability. Divergence may arise from the enhancement of regional differences in institutional capacities and socio-economic endowments, but convergence may prevail because of higher transparency in the allocation of resources and larger explicit equalization across jurisdictions. Most empirical studies have found that decentralization is associated with a general reduction in territorial disparities. This is an area where the evidence is quite robust. Decentralization generally contributes to regional economic convergence and more equal access to services (many of the references on these issues can be seen in Martinez-Vazquez et al. 2017). However, there is also heterogeneity in these trends, with results differing by country size and level of income. For instance, some research finds that low income countries observe positive associations between decentralization and convergence in access to basic services, in middle income countries convergence in quality of basic services, and in high income countries there is some evidence about convergence in income across territories. Nevertheless, as is the case in almost in all these relationships, there are other confounding factors that are difficult to separate from the effects of fiscal decentralization; for example, regional economic convergence may be after all the result of more general economic sources, such as decreasing marginal returns. Political and policy consequences of fiscal decentralization 20. There has also been developing interest in the impact of fiscal decentralization processes on political institutions and public policies. Here we select and briefly summarize the findings for a selected number of areas. Regarding the size of government or the public sector, decentralization is generally expected to contain it, as in the famed Leviathan hypothesis and Tiebout’s “voting with your feet” model. However, such impact may depend on the nature of decentralization and on how the public sector size is measured. When size is measured by the level of public employment, there is robust evidence that increases in sub-national employment following decentralization often overwhelms the corresponding decreases at the central level. This may be due to some form of inefficiency or, alternatively, simply a response to the expansion of labor-intensive public services, such as education and health, in decentralized systems. 17 21. In terms of the quality of governance, there is a mutually reinforcing relationship between fiscal decentralization and democratic governance. And governance is improved through increased accountability via elections and different forms of intergovernmental competition. On the other hand, and beyond theoretical positive and negative arguments regarding the impact on corruption, decentralization is commonly believed to foster public sector corruption. The big surprise is the solid empirical evidence finding that decentralization is robustly associated with both lower corruption and shadow economy. Decentralization also tends to encourage more collective action, interaction, and, ultimately, social capital and trust in government. 18 3. Fiscal decentralization: A Framework for the Assessment of Fiscal Decentralization System 22. Fiscal decentralization provides the link between incentives for better performance of the local government and the elected support from the citizens and is, therefore, essential for an effective system of decentralization. Fiscal decentralization is a set of rules that defines roles and responsibilities among different levels of governments for fiscal functions including budget preparation, budget execution, revenue generation, and public sector borrowing. Fiscal decentralization lies at the heart of any local government system as its rules define the generation and distribution of resources (both between and within different government levels) that are utilized to fulfil citizens’ demands. The ability of the government to make fiscal decisions in the provision of local government services is a precondition for the voters to assess the performance of their elected representatives with respect to the amounts and qualities of services they are getting for the taxes that they are paying (Mueller, 2006). Therefore, if subnational governments are denied the fiscal instruments and funding to make real use of their political and administrative autonomy, decentralization is likely to be ineffective. 23. This report uses a conceptual framework developed by the World Bank staff for the assessment of fiscal decentralization arrangements across client countries5. The framework itself is based on analytical work in the area of fiscal decentralization and the experiences of decentralization reforms from multiple countries. According to this framework, there are two primary components of a fiscal decentralization system: fiscal discretion and fiscal accountability. Fiscal discretion is granted to the local governments to perform fundamental fiscal functions and mechanisms that hold the local governments accountable for appropriate use of this discretion. Fiscal discretion can be divided into four elements, namely, the assignment of expenditure responsibilities, revenue generation, intergovernmental transfer systems, and local government borrowing. The framework emphasizes that an authority can be held accountable for performing a specific function only if it has the fiscal resources and the discretion to perform that function. Similarly, the framework argues that accountability is not an automatic outcome of increased discretion and that governments need to make a conscious effort to create structures that would hold local governments accountable. The two essential elements of accountability include public accountability, where responsible individuals are held accountable by other elected or non-elected officials, and social accountability, where public officials are answerable directly to the citizens. Figure 3 provides the graphical representation of these relationships. 24. Decentralization reforms in all dimensions are affected by political economy considerations. A comprehensive and detailed discussion of political economy is outside the scope of this report. The most important aspects of political economy of fiscal decentralization are introduced when relevant within the report. In addition, Annex 3 gives a broader perspective on political economy issues. 25. Further in this report, critical elements of such decentralization framework are described based on global good practices and particular experiences of decentralized countries. Responding to Chile’s demand, the emphasis is put on issues related to fiscal discretion, while the aspects of fiscal accountability are just briefly summarized. 5 World Bank (2010), How to note: A Framework for the Assessment of Fiscal Decentralization System, No. 123 Figure 3. Schematic illustration of the fiscal decentralization framework Expenditure Revenue Intergovernmental Subnational Borrowing and assignment autonomy transfers fiscal rules FISCAL DISCRETION FISCAL ACCOUNTABILITY Public Accountability Social Accountability Budget process Audit, control and Transparency Citizen participation evaluations Source: World Bank, http://siteresources.worldbank.org/EXTSOCIALDEVELOPMENT/Resources/244362- 1164107274725/Fiscal-Decentralization-web.pdf 20 4. Regional governments position in the decentralized government structure 4.1. Two vs. three-tier subnational government structure 26. As is to be expected the vertical structure of government varies from country to country. The diversity in the number of subnational tiers across different regions reflects a wide variety of reasons. There is literature identifying the main drivers of the vertical composition of subnational governments as well as the issue of jurisdictional size. The two most important determinants of the vertical structure of governments, that is, of the number of tiers of government, are simply population size and the land are of a country (Lago and Martinez-Vazquez 2013). Nevertheless, colonial legacies, geography or ethnic and linguistic fragmentation-- factors often summarized as the “long shadow” or the “dead hand” of history --- are also likely contributors to what vertical structure a country actually has. The average size of level of fragmentation of local governments (at the lowest tier) may also have an impact on the need for second or intermediate tier of government to supplement and support the delivery of some services for which small size local governments are not equipped to produce. 27. Globally, according to the OECD data, all countries in the sample had a municipal level of subnational government6. Among the OECD and Latin America countries, all countries have municipal tier of subnational government, while regional/state and intermediary levels have lower incidence, especially in smaller countries. Only seven OECD countries (Belgium, France, Germany, Poland, Spain, US and the UK) have an intermediate level of government. According to OECD, over recent decades, the intermediate level has lost many of its powers and responsibilities in favor of regions, that gained more importance. In a majority of countries, intermediate governments are now mainly responsible for administrative and delegated tasks and have small budgets and generally little to no taxing powers. When it comes to regional/state tier, only very small countries in the OECD group (Iceland, Israel, Estonia, Ireland, Latvia, Lithuania, Luxembourg and Slovenia) and within Latin America (Central America except for Nicaragua) do not have regional/state level of subnational governments. 6 This paragraph uses the OECD definition of subnational government tiers comprises a maximum of three levels of subnational governments: municipal, intermediate and regional. This structure facilitates country comparisons, but also hides a more complex organization in several countries, where several subnational government layers may exist. Within the same layer, several types of subnational governments can be found. For example, in Peru, within municipal level there are district and provincial municipalities which are sometimes considered two separate government tiers. According to the classification used in this report (after OECD), they are both considered “municipal tier”. Figure 4. Subnational governments tiers Source: Staff elaboration based on OECD-UCLG World Observatory on Subnational Government Finance and Investment (2020) 28. Fundamentally, fiscal decentralization does not mean that most or all public goods and services should be decentralized. Some public goods have such large benefit areas that they should be delivered by the central government. There are also economies of scale in production and large externalities in consumption that justify the production and delivery of other public goods at the regional level. And the subsidiarity principle indicates that many other public services can be assigned at the local level. Thus, depending on population and area size, an intermediate or regional level of government is justified in many country situations. In fact, there has been a significant trend worldwide in recent times to introduce ex novo regional level of government. The reasons for that are varied (See Box 1. Why are countries undertaking regional reforms?). Box 1. Why are countries undertaking regional reforms? Regionalization reforms, understood here as the creation of new regions, or the strengthening of existing ones, have represented an important trend in multi-level governance since the 1980s. Such reforms have been particularly important, with the influence of the European Union and cohesion policy, but other regions of the world, in particular Asia and America, have also experienced regional reforms in recent decades. The motivation for regionalization reforms varies across countries and often reflects individual geographic, social or political circumstances. Nevertheless, there are also some general characteristics that make regionalization reforms more/less likely in certain countries. For example, the size of the country matters: large countries tend to have more layers of subnational governments. But, many countries of a relatively modest size have also introduced or strengthened a regional level in recent decades. The main objectives behind regionalization reforms are an upscale in governance to generate countries of scale for public service provision, for example in the health or public transport sectors. Regions may also better respond to widening functional labor markets. Other objectives are to improve co-ordination between municipalities and intermediate levels of government in such areas as infrastructure delivery, spatial planning and land-use, for example. Larger regions are also expected to be more competitive: they have higher critical mass, more resources to implement effective regional development strategies, and the ability to foster intra-regional co-ordination and to implement more integrated territorial planning. Regionalization can have very different implications for decentralization: it can represent a form of decentralization with respect to central government, but it can also generate centralization at regional level with respect to local authorities; this situation is particularly common in the case of political regionalization and in federal states. In Europe, there have been in recent years trends towards 22 regionalization, that coexist with trends toward the reduction of the role of regions. Several countries have created new regions, notably in eastern European countries in the context of the EU enlargement process. For example, a regional governments tier was created in Poland in 1998 when the 49 former voivodeships were consolidated into 16 regional governments, which were better equipped to plan regional development in the context of large inflow of the EU structural funds. Other countries have strengthened existing regions: this is the case in recent or current reforms in Nordic countries, France, Greece or Italy. In several Nordic and Central and Eastern European countries, responsibilities such as higher education, specialized health care and regional public transport were reassigned from the municipal and the central government levels to a newly created regional level. In France, the 2015 reform merged and reduced the number of regions from 26 to 17 (of which 12 in mainland France, along with Corsica and four overseas regions). In Norway, it was decided to abolish the 18 counties and to replace them with 11 larger regions by 2020, in order to strengthen the regions as functional units and to provide more coherent housing and labor market areas. In Portugal, Lithuania and Bulgaria, the creation of self- governing regions is currently discussed. In Finland, it was planned to create self-governing regions, but the project was recently put on hold. In Sweden and Netherlands, regional mergers have been discussed but reforms were abandoned. However, in recent years some European countries have also gone in the opposite direction and have reduced the role of regions. This is the case in Denmark with a reduction in the responsibilities assigned to regions, for example, or in Hungary through recentralization reforms. In Estonia, the municipalities were recently strengthened by a radical merger reform, and at the same time the regional government level was abandoned, and the tasks were given to municipalities and central government agencies. In Asia, regionalization reforms have also been quite important, in particular since the 1990s. For example, in Japan, decentralization reforms were aimed to correct the excessive population concentration in Tokyo while in Korea, addressing the regional imbalance between Seoul and surrounding regions is also at the core of the decentralization program launched in 2017-18. Regionalization in Asia was often asymmetric, to recognize the specific identities of certain regions. For example, in the Philippines, as of 2019, the country comprises 17 administrative regions, with one being autonomous. In Latin America, regional reforms have also occurred in recent decades, but at a slower pace compared to other regions. In Peru for example, the current decentralization process started in 2002, when Congress, driven by democratic and economic objectives, constitutionally declared Peru a “decentralized state” (Article 43 of the constitution). In Colombia, the adoption of the new constitution in 1991 significantly enhanced the process of decentralization, declaring that “Colombia is a legally organized social state under the form of a unitary, decentralized Republic with autonomous regional entities.” In Chile, administrative regions became self-governing entities in 2017 and the first regional elections are to take place in 2020. Source: OECD/UCLG (2019), 2019 Report of the World Observatory on Subnational Government Finance and Investment – Key Findings 4.2. Hierarchical dependency of subnational governments 29. In the international experience (including developed and developing countries, with constitutional federal and unitary regimes) there are two main models of fiscal interrelationships among the different levels or tiers of government: (i) the hierarchical model, and (ii) the bifurcated model. Many countries use a combination of both designs, but most often with the predominance of one of the models over the other. 30. In the case of the hierarchical model, the central or federal government interacts exclusively with the intermediate level of government (provinces, regions or states) and never, or almost never, does so directly with the lower level of government (local, municipal or district governments). In the hierarchical 23 model it is only the regional governments that interact directly with the local governments located within its boundaries. Most, if not all policies of the central authorities that may have ultimately a local dimension of impacts go through the intermediate level of government. These hierarchical flows extend to most, or all, aspects of intergovernmental fiscal relations, form functional coordination, to revenue sharing, or the design of equalization grants and conditional grant programs. 31. In contrast, in the case of the bifurcated model, there is no direct hierarchical relationship between the intermediate or regional level of government and the lower tier of local governments. In this case, the central government implements separate systems of relations with the regional governments and the local governments. And all these bifurcated flows extend to most, if not all, aspects of intergovernmental fiscal relations, from functional coordination, to revenue sharing, or the design of equalization grants, conditional grant programs, and so on. 32. The presence of the hierarchical model is common in federal countries, because most federal constitutions explicitly recognize the powers emanating from the states and federal government, while local governments are either not explicitly mentioned or are simply considered to be dependent on the states. 7 The hierarchical system is also frequently found in the case of unitary countries, sometimes with past histories of strong centralization, but not necessarily always so. Thus, decentralizations systems as diverse as those of the United States and China can be considered hierarchical. The presence of the bifurcated systems is most common in the case of loosely defined unitary countries, with varied examples such as Poland, Spain and the United Kingdom in European Union and Ukraine in Eastern Europe.8 4.3. Types of decentralization and de-concentration versus decentralization in parallel structures 33. Decentralization – the transfer of authority and responsibility for public functions from the central government to subnational governments – is a multi-faceted concept encompassing many different blends of political, administrative and fiscal decentralization. These dimensions are inter-dependent: there can (or should) be no fiscal decentralization without political and administrative decentralization. On the other hand, without fiscal decentralization, political and administrative decentralization are meaningless. Drawing distinctions between these various concepts is useful for highlighting the many dimensions of decentralization (Figure 5). 7One exception of a federation with a bifurcated system is Brazil. 8In the case of Ukraine and other countries in transition the argument has been made that a bifurcated system was seen as a solution to past histories of abuse or neglect of local government by regional governments. 24 Figure 5. Political, administrative and fiscal decentralization Source: OECD (2019) 34. Political decentralization empowers citizens and their elected representatives to make decisions in public affairs that directly affect their lives. While it is often associated with pluralistic policies and representative government, it can also support democratization by giving citizens greater influence in the formulation and implementation of policies. Effective political decentralization may require constitutional or statutory reforms, strengthening legislatures, creation of local political units, and encouragement of citizen participation. 35. Administrative decentralization enables subnational governments to run their own government affairs separately from the organs of the central government, making decisions on the best way to produce and deliver services. The most important component of this type of decentralization is the ability of subnational governments to determine the salaries and general compensation of their employees and to hire and fire personnel as they consider it appropriate in carrying out their functional responsibilities. 36. Fiscal decentralization aims to redistribute authority, responsibility, and financial resources for providing public services from the central government to subnational governments. Administrative decentralization differs from fiscal decentralization in that it entails the transfer of administrative functions from the central government to agencies, field units of government agencies, subordinate units of government, or area-wide, regional or functional authorities without transferring financial responsibility. Three distinctions are made in the literature: 37. Deconcentration is not, strictly speaking, a form of decentralization. It involves the redistribution of decision-making authority and financial and management responsibilities for service delivery from the central government to regional or local offices of the central government, thereby, preserving the hierarchical relationship between field staff and the central government. Under deconcentration, local decision makers respond and are accountable to central authorities and not to local constituencies. 38. Delegation involves the transfer of limited responsibility for decision-making and administrative functions from the central government to semi-autonomous subnational entities not entirely controlled by the central government, but ultimately accountable to it. It is a “top -down” process by which the central government gives subnational governments the power to provide services and to raise resources according to explicit norms and rules set by the central authorities, thereby circumscribing the authority or autonomy of subnational governments. Such a form of decentralization is most often associated with unitary forms of government. 39. Devolution entails the transfer of authority for decision making, finance and management to quasi- autonomous subnational governments, which have legally recognized geographical boundaries over which 25 they exercise authority and within which they perform public functions. This is best described as a “bottom- up” process as subnational governments have a more permanent right, such as a constitutional right, to govern their own affairs, including the ability to raise taxes and formulate their expenditure budgets. This approach to decentralization is sometimes identified with federalist forms of government and may result from devolution of legislative authority to members of a federation operated under the Constitution. It may also result, in both unitary states and federations, from a transfer of executive authority (by law) to subnational governments. 40. In many decentralized systems it is possible and desirable to observe the coexistence of all three forms of decentralization. While devolution together with delegation represent the core activities of decentralized subnational units, deconcentration may still be present in terms of the territorial operations of central or federal government line ministries and other agencies. The presence and coexistence of decentralized subnational units and deconcentrated central government units for the provision of separate well identified services is a model that provides more transparency and accountability, and therefore overall higher efficiency, than models where the lines of decentralized and deconcentrated functions are blurred and provided by decentralized units. 26 5. Fiscal discretion 5.1. Assignment of expenditure responsibilities 41. Expenditure assignments answer the fundamental question of which level of government – local, regional or central – should be responsible for the specific functions and activities that do belong and will remain in the public sector. 42. The first fundamental step in the design of a system of intergovernmental fiscal relations should be a clear assignment of functional responsibilities among different levels of government. The international experience clearly shows that instability, controversy, and ultimately inefficiency in public service delivery follows when the decentralization laws are silent or unclear about the competencies and expenditure obligations of different levels of government. Designing the other important pieces of a system of decentralized finances, revenue assignments and transfers, in the absence of a clear expenditure assignment (as has been a common experience in Latin America, (e.g. Peru) and many countries in transition) is to put the cart before the horse. 43. A previous necessary step to expenditure assignment is to neatly define the roles of the private and public sectors in the economy. The role of the public sector should be limited only to those goods and services that private markets will not supply efficiently - either because of difficulty of charging an exclusionary price or because of joint-consumption benefits - or those which have significant re- distributional value. Only then can the question be addressed of which level of government - local, regional or central - should be responsible for the specific functions and activities that need to be provided within the public sector. 5.1.1. General principles of expenditure assignments and international experience 44. The international experience (including developed and developing countries, with constitutional federal and unitary regimes) shows great diversity in expenditure assignments, which reflects the different approaches to decentralization and other relevant factors, including the federal versus unitary nature of the state, history, geography or level of development. 45. Expenditure assignments in most countries are broadly aligned with the principles of expenditure assignment. The most salient of these principles is the “correspondence” principle: the geographical dimension of the benefits received from a particular service should match the geographical dimension of the level of government responsible for its provision. Thus, services spreading throughout the national territory (for example, air traffic control) should be assigned at the central level, services with benefits with regional dimensions (for example, tertiary specialized hospitals) at the regional government level, and those with a local dimension (for example, street cleaning) at the local government level. 46. Another important principle is the “subsidiarity” principle, which means that the responsibility for any service that can be provided at different level should be assigned at the lowest possible level of government compatible with the size of “benefit area” associated with those services.9 9The principle is now widely accepted. For example, the European Chart of Subnational Governments formally adopts the “subsidiarity” principle as the most prominent guide for expenditure assignments. 47. Efficient service provision may be enhanced by making use of the “benefit” principle, by using fees and service charges from users of the service to recover the costs of provision. These “pricing” mechanisms help reveal residents’ demands and preferences to local officials and may also reduce usage congestion by rationing demand. However, fees and charges are not so frequently used because of fairness considerations regarding lower income residents. 48. Last, and to an extent already implied by the correspondence principle, service spending programs openly pursuing distributional and macro stability objectives should be reserved for the central government level. Local and regional governments may have difficulties sustaining independent (self-financed) redistributions programs because they risk attracting the needy from other areas while taxing more heavily their (potentially mobile) residents. For stabilization policies, such as massive investment or unemployment compensation, subnational governments may be relatively too small to be effective and their efforts may be quickly diluted within the open borders with the rest of the national economy. 49. It is important to note that the application of the principles above facilitates the assignment of expenditure responsibilities to different levels of government but does not always yield an unequivocal answer. Some public services, e.g., primary education and primary health services, may be of a local nature by the size of their benefit area, but because of their relevance in welfare and income redistribution they may also be considered a responsibility of the central government. It is in this sense that it is not meaningful to talk about the best assignment of expenditure responsibilities. A fair conclusion is that “there is no single best assignment.” What may be considered a good assignment is likely to change over time with changes in the costs of provision as well as technological innovations and changes in citizens’ preferences. Nevertheless, there is a need at any given moment in time to have a concrete, clear and stable assignment of expenditure responsibilities. 50. Defining expenditure assignment also requires considering issues related to affordability and capacity: • Affordability: has to do with the extent to which subnational expenditure needs can be covered; that is whether or not subnational authorities have the sufficient resources to adequately deliver the service and expenditure responsibilities assigned to them. Although this is a basic question, it is also a difficult one to answer well. The reason is that the measurement of expenditure needs arising from any particular expenditure assignment is a complex and inherently ambiguous task. Several approaches are used in practice for the measurement of expenditure needs (which are discussed below in relation to the design of equalization transfers). The important issue here is to emphasize that any and all assignments of expenditure responsibilities derive necessarily into certain levels of expenditure needs. Whether or not subnational governments have enough resources to provide some minimum services is typically a complex issue; the design of the financing system for subnational governments (how much they should raise in own revenues, what transfers they should receive, and so on) in large part attempts to address this question of financing adequacy. • Capacity: is important to determine whether subnational governments are administratively capable and politically responsible to have a good sense of duly prioritizing their resources to provide adequate levels of services, especially in those that may be considered high priority sectors, such as health and education. This tends to be a thorny issue especially in newly decentralized systems where there is mistrust about the capacity and qualification of newly appointed subnational authorities. This is an issue that may drive the categorization of expenditure responsibilities in obligatory versus discretional, as discussed further below, on whether compulsory minimum standards of services for subnational governments are introduced, and on how the system of transfers may be designed, for example, by using specific transfers versus more general block grant or unconditional transfers. 28 Box 2. Most Common Problems with Expenditure Assignments The international experience also provides some valuable lessons on what to avoid as poor practice. • The most commonly found problem is the lack of a formal explicit assignment of responsibilities. Although it is true that highly decentralized countries, such as the United States, may lack a formal assignment of expenditure responsibilities, these countries have a de facto assignment that has been achieved through many trials and tribulations during very long periods of time, in some cases centuries. Countries that are newly decentralizing are better served by legislating an explicit assignment of responsibilities; otherwise they risk a contentious and inefficient process of sorting out responsibilities in practice for decades to come. • A second common problem in the assignment of expenditure responsibilities is the presence of the wrong or inefficient assignments, which openly brake with the principles above. This is the case, for example, in the assignment of expenditure responsibilities, including financing, for social protection and welfare to local governments, or the assignment of local fire protection services, or water and sewerage services at the central level, or the assignment of all capital expenditure responsibilities at the central level (an issue further discussed below). • A third common problem is the extensive use of concurrent responsibilities (those assigned simultaneously to two or more level of government) without clarifying the level of government responsible for the different “attributes” of the service: regulation, financing and implementation of those services. This is often the most common cause of confusion, friction and inefficiencies. Ideally, each service should be assigned exclusively to one single level of government, so that the line of accountability is transparent. In reality some concurrent responsibilities are inescapable because different levels of government have advantages regarding certain aspects or attributes of the service. Overall, concurrency of services becomes less of a problem when those particular attributes common to any service (financing, regulation, and implementation) are clearly assigned to different levels of government. Source: Own elaboration 5.1.2. International experience in expenditure assignments 51. Expenditure assignments at the national, regional and local levels differ significantly across countries. Box 3 offers a representative assignment of expenditure responsibilities for a typical list of government functions. The assignment shows which level of government (national, regional or local) could properly be involved in the regulation, financing, provision or production of these government functions. The summary confirms that indeed there is no single best assignment of responsibilities. For instance, the responsibility for secondary education may properly be placed at the regional level in some countries, whereas in other countries and under other conditions secondary education could be appropriately considered a local government function. The summary further reveals that many (if not most) government services ultimately will involve more than one level of government. The fact that the devolution of expenditure functions often involves several levels of government emphasizes the need for intergovernmental cooperation and dialog in order to assure the successful delivery of public services. 29 Box 3. Representative Assignments of Expenditure Responsibilities10 Provision / Function Regulatory Financing Product. Administration International affairs N N N N Defense N N N N Public order and safety N,R,L N,R,L N,R,L N,R,L Primary and secondary N,R,L N,R,L N,R,L N, R,L,P education Higher education N,R N,R N,R N,R,P Health care N,R,L N,R,L R,L R,L,P Social security and welfare N,R N,R,L R,L R,L Community services (water, N,L R,L R,L L,P sewer, refuse, fire protection) Highways, roads and streets N,R,L N,R,L N,R,L N,R,L,P Parks, recreation and culture N,R,L N,R,L N,R,L N,R,L,P National transportation / N N,R N,R,L N,R,L,P communication networks Regional/local public R,L R,L R,L R,L,P transportation Note: N denotes National Government; R denotes Regional Government; L denotes Local Government, and P denotes the private, non-government sector / civil society. Source: Own elaboration summarizing multiple sources 52. Services provided by regional level of governments also differ significantly across countries. Box 4 provides examples of the most typical assignment of responsibilities to this level of subnational governments. Annex 2 provides a comparative table for Latin America countries with description of key expenditure responsibilities at the regional/state level. Box 4. Breakdown of responsibilities at the INTERMEDIARY and REGIONAL level1 Specialized and more limited responsibilities of supra-municipal interest, with an important role of assistance towards small municipalities May exercise responsibilities delegated by central government Responsibilities determined by the functional level and the geographic area: for example: • Secondary or specialized education • Supra-municipal social and youth welfare • Secondary hospitals • Waste collection and treatment • Secondary roads and public transport • Environment Heterogeneous and more or less extensive responsibilities depending on countries (in particular, federal vs unitary): Services of regional interest, for example: 10The information in this box is stylized representation of the common threads found in most countries (developing and developed) across the world with a decentralization system of government and covering decades of actual practice. For specific sources see World Bank (2004, 2008 and 2009), Shah (1991); Martinez-Vazquez (1999); McLure and Martinez-Vazquez. (2000); Martinez-Vazquez and Timofeev (2007). • Secondary / higher education and professional training • Spatial planning • Regional economic development and innovation • Health (secondary care and hospitals) • Social affairs e.g. employment services, training, inclusion, support to special groups, etc. • Regional roads and public transport • Culture, heritage and tourism • Environmental protection • Social housing • Public order and safety (e.g. regional police, civil protection) • Local government supervision (more often in federal countries). Source: OECD/UCLG (2019) 5.1.3. Addressing the assignment of capital investment responsibilities 53. Subnational governments play a key role in public investment. As recently reported in OECD/UCLG (2019), the level of subnational government investment exceeds 40% of total public investment in high-income economies and subnational governments are important investors across all geographic regions. Oftentimes, the high ratio of capital expenditures at subnational level is, in part, due to the legal constraints that make it mandatory for local governments to spend a percentage of subnational budgets in capital expenditure (often through revenue earmarking and special capital transfers). Figure 6. Subnational government investment as a share of public investment by income groups and geographical areas (2016) Source: OECD (2019) 54. Typically, the assignment of expenditure responsibilities for service delivery in a recurrent basis and the responsibility for building the infrastructure required for the provision of the services may be assigned at different levels of government. This may be the case because infrastructure financing may be beyond the means and the ability of the level of government responsible for the provision of recurrent services. 55. However, the international experience (most dramatically evidenced in countries in transition from socialism) also shows that splitting the responsibilities for recurrent activities and capital infrastructure investment for that same service does not always lead to good results. One main problem is that the level of government responsible for current services (typically a local or regional government) is also in charge of not only, logically, operating the infrastructure facility, but also in charge of its maintenance. What may follow is a significant lack of infrastructure maintenance expenditures or generally a not sustainable use of the infrastructure facilities by the subnational government (perhaps with the hope that the upper level government that constructed the facility will eventually repair it or replace it). Beyond the perverse 31 incentives that are built in the splitting of responsibilities, subnational governments may lack the funds for the adequate maintenance. As it is well known, infrastructure maintenance is one of the areas that governments tend to postpone when facing financing problems. 56. Therefore, best practice is to assign both recurrent and investment responsibilities for any particular expenditure function at the same level of government. This solution is clear in the case of exclusive assignments, even though it will require making provisions on the financing side of the decentralization system for financing capital expenditure needs of subnational governments, which generally speaking will require some combination of capital transfers and borrowing. The solution is more complex in the case of concurrent responsibilities. What is desirable in this case is to avoid the splitting of financing responsibilities for recurrent and investment obligations among different levels of government, while concurrence may exist in other attributes or obligations, such as regulations or actual implementation and delivery. 57. As recognized in the OECD (2019), effective co-ordination of public investment among levels of government is crucial, especially in cases when it is a shared responsibility across levels of government. Adequate multi-level governance arrangements are crucial to identifying investment opportunities and bottlenecks, managing joint policy competencies, ensuring adequate resources and sufficient capacity to undertake investment, resolving conflicts, or creating trust. Recognizing the importance of these challenges, the OECD Regional Development Policy Committee developed the Recommendation of the Council on Effective Public Investment Across Levels of Government, adopted in 2014. The Recommendation groups 12 principles into three pillars to help governments at all levels assess the strengths and weaknesses of their public investment capacity and set priorities for improvement: (i) Co-ordinate across governments and policy areas (ii) Strengthen capacities for public investment and promote learning across levels of government; and (iii) Ensure sound framework conditions at all levels of government. 5.1.4. How to go about assigning responsibilities to regional governments? 58. First, several decisions need to be made prior to the reform of expenditure responsibilities, which are discussed here. These steps and decisions do not correspond to the particular experience of any single country, but rather they summarize best practices observed in many different countries, developing and developed, and also countries in transition that during the decade of the 1990s struggled with reconverting the role of the state omnipresent in planned socialism systems to that in a market economy. The relevance of the different steps to any particular country is dependent on its history with decentralization and the role of the public sector in the economy: • Define the scope of the reform in expenditure assignments and privatization of those private-in- nature activities performed by government (if any): This may be obvious, but the reform of expenditure assignments must cover all government activities, including budgetary functions as well as extra-budgetary functions and regulatory powers. • Examine the effectiveness and efficiency of the current assignment of expenditures. For a country embarking on fiscal decentralization reform in the context of creating a new layer of subnational government, it might be relevant to take stock of the hitherto assignment of expenditure responsibilities and assess the quality and efficiency of service delivery. There is no need to reinvent the wheel when things are working well. Decentralization reforms are supposed to improve and enhance delivery of service and improve the quality mainly in areas where significant gains can be expected, i.e. where the current delivery model may not be optional. • Decide to use general assignments versus different types: In determining expenditure assignments 32 we may want to make a distinction between voluntarily (or “own” responsibilities) versus mandatory functions of subnational governments. That is, to differentiate between what sub- national governments can do and what they must do. And in the latter category, whether they are “delegated” mandated responsibilities (the central government is ultimately responsible but delegates implementation on subnational units) or “own devolved” mandatory responsibilities (to be financed locally). Note that this is not a requirement for a clear assignment of expenditure responsibilities. Typically, when the distinction is made between delegated and own responsibilities, the central government also assumes the responsibility for directly financing in some form those functions that are delegated to sub-national governments. More frequently, the arrangement for paying for delegated responsibilities is via conditional or tied grants. In these cases, sub-national governments need to spend the funds in the way specified by the central authorities.11 • How far to carry “exclusive” (as opposed to “concurrent”) assignments : As already discussed, exclusive responsibilities are more conducive to accountability and efficient provision; however, in practice, we have also seen that exclusive responsibilities are not always possible and in some cases they may not be even desirable.12 In some cases, under concurrent responsibilities, the national government “regulates” the service by establishing standards in the form of some “framework legislation,” while sub-national governments can decide on the details and deliver the services (e.g., Germany, Spain, Mexico, Brazil). • Decide on the general approach to use for the assignments: whether to a dopt a “general competence” principle or a “closed list” for defining the assignment of expenditure responsibilities: A closed list approach is more cumbersome, but it can offer more protection to sub-national governments against encroachment by central or upper-level governments. Typically, the question boils down to what level of government may have “residual powers” in case a particular (new) function does not appear in the law in the closed list of any one level of government. For example, the European Charter of Local Self-Government provides “residual powers” to local governments. On the other hand, it may be easier to assign “general competences” and list a certain number of exclusions. The actual practice with this issue is varied, frequently with historical roots. In the case of federal systems, a closed list of national functions with an open list for subnational units is common in federations, especially in those established as a union of previously distinct states. Residual powers also remain with the sub-national units. Unitary countries and federations or quasi-federations that have evolved from formerly unitary countries (or unitary regimes) tend to have a closed list for subnational governments and an open list plus residual powers for the national government. • Choose the legal instrument employed to specify expenditure assignments: Because assignments should be stable over time, they should be legislated into law, preferably a higher standing or organic law such as a Decentralization Law or the Budget Code, which cannot be nullified by other (posterior or existing) ordinary laws. It is generally less preferable to introduce expenditure assignments in the Constitution because of the detail required and the difficulty of changing this fundamental law. A balance is needed between flexibility and rigidity of expenditure assignment because changes will be needed in the future, due to new technologies and so on; in that sense a special organic law may be a good choice. It is often the case, that sector specific laws (law on 11 It is worth pointing out that although the practice of conditional or tied grants from the central to subnational governments is practically universal, the majority of countries implementing this type of grants do not make the distinction between “delega ted” and “own” responsibilities of sub-national governments. 12 Note, however, that some countries, such as Germany, have adopted the model “cooperative federalism,” in which most assignments are concurrent and different levels of government continuously interact with each other via different committees in the design and implementation of public service delivery. 33 education, law on the health system, etc.) make also reference to the assignment of functional responsibilities in the respective sectors. In this case, becomes extremely important to harmonize any existing sectoral laws with the organic law addressing the expenditure assignments in general. • Decide on the level of detail to be specified in the law versus how much can be left to coordination mechanisms or ultimately to the interpretation by the courts in the case of disputes: It is clear that no expenditure assignment can be ever detailed enough to preempt the need for dialog and coordination among different levels of government. In practice this decision is influenced by the type of legal system the country has. Countries operating under a common law tradition tend to rely more on precedent practice and judicial review, while countries with a codified civil law tradition tend to rely more on detailed laws and regulations for the implementation of governmental functions. • Decide on having uniform versus asymmetric responsibilities: Although long existing decentralized systems commonly have uniform assignments, there has been a growing trend, especially among newly decentralized countries, to adopt asymmetric assignments across subnational units. One of the most vexing problems newly decentralized systems face is the small size of traditional local jurisdictions, too small to realize economies of scale in the delivery of public services or lacking adequate fiscal and managerial capacity. Although there are some ways to get around the small scale such as privatization of services, the creation of special districts, or associations of local governments, the most preferable solution to the problem is the amalgamation of smaller local governments into larger units with adequate size and capacity. Unfortunately, this solution (amalgamation) is often politically not feasible. It is here where the asymmetric assignment of expenditure responsibilities adjusted to the capacity of local governments also appears as a solution. In theory at least, this type of asymmetric treatments of jurisdictions should be temporary because local governments will have an incentive to use all the technical assistance provided so that they can graduate and assume the competencies assigned to their peers. A different type of asymmetric assignments is more political and tend to recognize the special status for historical and other reasons of some units. This type of asymmetry tends to be long lasting and it can help subdue centrifugal forces but also can create resentment among other “non-privileged” regional governments. • Who should be in charge of these tasks? Leaving aside the historical federal systems in which the constituent states “delegated” a closed list of powers to the federal governments, the international norm is that the determination of expenditure assignments is the prerogative of the national or central authorities. This is true of unitary countries and also a number of federal or quasi federal countries. The reform of expenditure assignments needs to be a comprehensive exercise involving many central government agencies as well as representatives of subnational governments. It should involve consultations with line ministries and other central government agencies. Expenditure assignment reform can involve significant transfers of personnel and assets to subnational governments and the necessary refocusing of central line ministries and agencies from the direct delivery of services to regulation and enabling activities in support of the decentralized governments. None of these activities will go smoothly and efficiently without significant buy-in of those agencies. Consultations with and direct involvement of subnational governments in the reform process via their associations and representatives is also important. • Significant data and other information will be needed: Setting up to conduct comprehensive reform of expenditure assignments will require significant amounts of data on actual expenditures and potential expenditures or expenditure needs associated with the reassignment. Getting all this information on a timely basis may be one of the most significant hurdles to expenditure assignment 34 reform. Significant efforts are also typically needed for the review of contradictions and incompatibilities between the expenditure assignments legislation and existing sectoral laws and other legislation, existence of unfunded mandates, and so on. • The monitoring and enforcement of the new assignments: In practice, in many countries there can be significant differences between what the assignments are in practice and what is specified in the law and regulations. Thus, subnational governments may be providing services that are assigned at the central level but that central authorities may fail to deliver properly, or central government agencies may continue to provide services and exercise functions that have been long decentralized. Lack of compliance with formal assignments may worsen with comprehensive reform of expenditure assignments because of resistance to change and also out of inertia and lack of information. These problems will tend to be minimized if the process of reform is highly participatory involving all stake holders and the reforms are well publicized. The added clarity to the assignments of expenditure responsibilities by the reform should also contribute to reducing differences between practice and law. • Intergovernmental institutions for cooperation and dialog among the different levels of government can also be of great help. This should involve regular and transparent exchanges of information and dialogue at the political and technical levels. At the political level, there should be periodic meetings between the executives of different levels. Similar periodic meetings should be held in each of the main sectors where there are concurrent responsibilities such as education or health. Countries that have active intergovernmental committees to inform and discuss all types of issues are countries that tend to show less conflict and non-compliance with formal expenditure assignments. 59. Practical Steps for reforming/ establishing expenditure assignments are included in Annex 1. 5.2. Generation of income and revenue autonomy 60. Revenue autonomy is a complex issue and goes beyond tax autonomy. At the same time, revenue autonomy is at the core of fiscal decentralization. The proportion of grants versus tax revenues in subnational revenue and the power on these sources of revenue is key. The benefits of greater revenue autonomy are very important in terms of political accountability, expenditure efficiency, fiscal and budget responsibility, and the hardening of the budget constraint. It is important to underline that there are no clear- cut frontiers between the different sources of revenue. Rather, it is more of a continuum with fuzzy delimitations, ranging from very little autonomy for earmarked grants (allocated for specific tasks or projects and coming with guidelines, stricter controls and reporting obligations) to high autonomy for own- source taxes when subnational governments have the power to set rates and bases (e.g. typically property taxes). Box 5. Revenue autonomy Subnational fiscal autonomy is the extent to which Subnational Governments (SNG) rely on locally raised revenues for funding (rather than receiving transfers from central government) and on their ability to set their own tax rates. Thus, a truly local tax is one for which the local government (LG) can: • decide whether to levy the tax or not; • determine the precise base of the tax; • set the tax rate; • administer (assess, collect, enforce) the tax; • keep all the revenue collected; 35 • grant tax allowances or reliefs to individuals and firms The OECD has established indicators that rank local fiscal autonomy for LGs ranging from full power over tax rates and tax bases at one extreme to no power over rates and bases at the other extreme. Tax sharing, whereby the central government collects revenue from a tax and shares it with LGs, appears entails little or no local autonomy because the LG has no control over the tax rates or tax base. Tax sharing is virtually almost the same as an intergovernmental transfer. Source: Adapted from Making Decentralization Work: a Handbook for Policy-Makers, OECD, November 2018. 5.2.1. Own revenue sources 61. Is there really a need for assigning regional governments their own tax revenues? Why not simply finance them with different types of central governments transfers? 62. The fundamental premise of fiscal decentralization is that it cannot only strengthen democratic institutions, but also increase the overall efficiency of the public sector and lead to improved social and economic welfare. One critical assumption behind those results is that decentralized governments will generally be more accountable and responsive to citizens’ needs and preferences. At the same time, there is general agreement among experts in decentralization that the increased accountability associated with decentralization can only be assured when subnational governments have an adequate level of autonomy and discretion in raising their own revenues. The logic is simple but powerful: subnational authorities will spend resources differently when they need to incur the political costs of taxing their constituents, in comparison to what they would do if the money comes to them, practically cost free, from the central government. This is the argument behind the “tragedy of the commons.” When subnational expenditures are being financed by other taxpayers outside the jurisdiction, there will be perverse incentives to overspend and spend poorly. 63. But, if effective fiscal decentralization requires meaningful revenue autonomy at the regional and local levels of government, then the important question is which taxes should be allocated at these levels? This is what is known as the “tax assignment problem.” A relatively common mistake repeated in processes 36 of decentralization over the past several decades in many countries — and the Latin America region is a prime example — has been to take lightly the need for revenue autonomy and just use instead weak forms of revenue decentralization, such as revenue sharing. Often times, this has led not only to failing to capture the benefits of decentralization, but also to different forms of fiscal irresponsibility, sustained aggregate budget deficits and macroeconomic instability. 64. Fiscal decentralization reform needs to follow an explicit sequence; the known dictum of “finance should follow function” means that that revenue assignments should only come after the assignment of expenditure responsibilities, and the associated expenditure needs, has been completed. Prior knowledge of expenditure assignments can also help to better design revenue assignments because different services may call for different forms of financing. For example, some services (public utilities, bus transportation) can be financed by user charges while other services characterized by significant externalities, should be financed from region-wide taxes, while delegated function should be financed with conditional grants. 65. Despite its benefits, in the international practice we observe, with some exceptions, low levels of tax autonomy; but, why? From a political economy perspective, that outcome can be interpreted as a perverse equilibrium, which is desired by the two main players involved. Central governments are reluctant to devolve taxing powers for fear of having to compete with local governments for the same tax bases and/or fear of losing control of fiscal policy. At the same time, subnational governments are reluctant to take on the responsibility of making politically unpopular decisions to raise their own taxes. Thus, from the start, careful and willful design for revenue assignments is paramount. 1.1.1.1. Subnational tax autonomy 66. Even if regional tax autonomy is needed, not all forms of tax autonomy are desirable. There are multiple trade-offs in designing the tax structure at the subnational level. The most important aspects are highlighted below. 67. Open versus closed lists of subnational taxes: There are several main dimensions of tax autonomy. The first is which level of government has the right to choose the taxes that this given level can impose. There are good reasons to limit the ability of sub-national governments in this form of autonomy.13 The general choice is between an open list of taxes to be determined by the subnational governments themselves within some general limits and restrictions, or instead a closed list of allowed taxes which is determined at the national level, and from which subnational governments can choose. Overall, a closed list of sub- national taxes is arguably preferable because it avoids the introduction of nuisance taxes in some cases or higher and inefficient distortionary taxes which can impede local economic development, domestic trade and growth. In practice, where sub-national governments are given more constitutional discretion, as in the case of some federal systems, open lists with some general restrictions are common. Closed lists are used more frequently in unitary systems of government. 68. Co-habitation or exclusiveness of tax bases: Whether an open list or closed list approach is adopted an additional decision needs to be made as to whether the base of specific taxes should be used exclusively by one level of government, or whether these bases can be used simultaneously (co-habited) by several levels of government. In practice, the choice between exclusive or shared tax bases among tiers of government comes down to weighing the advantages and disadvantages associated with each choice. The main disadvantage of cohabitation is the potential presence of negative vertical externalities (tax decisions at one level affect the tax bases and revenues of other levels of government). The most important disadvantage of using the exclusive basis is that typically subnational governments will be shut out of any 13 For example, the Commerce Clause in the United States’ Constitution prohibits the use of internal tariffs by the states 37 opportunity to use significant (either in size or over time buoyancy) tax bases, thus drastically reducing any meaningful possibility of sub-national tax autonomy. The imposition of exclusive tax bases can also lead to cumbersome tax structures, such as dividing the VAT base between goods and services in India and Pakistan, which interferes with the efficiency of the tax. All things considered, the preferred approach may be a choice of a closed list allowing for the co-habitation of tax bases by different levels of government and possibly using intergovernmental transfers to correct for any potential vertical externalities. 69. Autonomy over what aspects of the tax structure: Another dimension of tax autonomy relates to which level of government can legislate over the structure of the tax bases and which level has discretion to set the tax rates. Of these two types of autonomy for structuring sub-national taxes, autonomy to define the tax base is generally less desirable than autonomy to set tax rates. 14 Variations in the definition of the tax base, either through especial exclusions from tax, deductions from the tax base, and credits against the tax liabilities can more easily lead to complexity and lack of harmonization across jurisdictions. 70. The importance of tax harmonization among subnational taxes: The most important unwanted consequence of the lack of harmonization and the complexity that follows is the higher tax administration cost for all the jurisdictions involved, as well as the higher compliance costs for taxpayers who have tax obligations in several jurisdictions. On the other hand, autonomy to define the tax rate generally tends to be more desirable because it is simpler to deal with across jurisdictions for both tax administrators and taxpayers. Focusing on autonomy in tax rate setting has the additional important advantage of enhancing the visibility of the tax, which tends to generate greater political accountability. Both households and businesses have an easier time figuring out the fiscal exchange or net benefits provided by different jurisdictions in their tax-public service packages when the differences in tax burdens are expressed in terms of (highly visible) differences in tax rates. The achievement of tax harmonization is easiest when tax the tax structure for subnational taxes, including the definition of tax bases, is nationally legislated, still leaving discretion on rates for subnational governments to decide actual burdens.15 The harmonization of subnational taxes when tax structures are separately legislated by subnational government can prove to be quite difficult. 71. Who should administer subnational taxes: The last dimension of revenue autonomy refers to which level of government is put in charge of administering various taxes. That a particular tax has been assigned at the local level does not necessarily mean that it should be administered at that level; under some circumstances it may be more advantageous to have that tax centrally administered, but with the subnational government still making the policy decisions of setting tax rates. The international experience shows a variety of approaches to the organization and degree of decentralization in tax administration. Countries with considerable decentralized revenue authority may have highly centralized tax administration (e.g., Scandinavian countries) and countries with little decentralized tax autonomy may have highly decentralized tax administration (e.g., Germany). Outside those polar cases, one can find systems with separate tax administrations (each level of government administers its own taxes) or mixed models (the central 14 The ability to change either base or rate opens up the possibility of fiscal competition among sub-national governments. Inter- jurisdictional fiscal competition can have both good aspects, such as offering choices to taxpayers and keeping public officials more accountable, and also bad aspects, such as a “run to the bottom” type of behavior actually taking place in countries with a significant degree of sub-national tax autonomy. In addition, the ability to change tax base or rate can give rise to “horizontal” fiscal externalities, whereby the policies of one jurisdiction (for example, raising tax rates) can have an effect on the tax bases of other jurisdictions (raising their tax bases related to mobile taxpayers). Intergovernmental grants can also be implemented by the central government to correct horizontal fiscal externalities. 15 One strong form of ensuring subnational tax harmo nization is to use the figure of “ceded taxes’ (as in the case of Spain’s impuestos cedidos), whereby the subnational tax is a part of the national tax, thus having the same tax base definition, but leaving discretion to the subnational governments to change the rates. The main disadvantage of this approach is the lower visibility of the subnational tax to citizens, who may continue to perceive the tax as a national tax, therefore leading to lower levels of accountability. 38 government administers some local taxes, and much less frequently local governments administer some central taxes.) 72. From a technical side several factors affect the choice of centralized versus decentralized structure. Some favor a more centralized administration, including: economies of scale and scope, informational externalities, and compliance costs with uniformity of tax procedures. One in particular favors a more decentralized mode: accountability to the residents. In addition, there are political economy factors such as opportunities for corruption and the creation or control of public employment that can become relevant. A primary advantage of centralized tax administrations is the ability to operate with lower costs through a more efficient use of inputs because of economies of scale in production, greater specialization of staff, and more sophisticated uses of capital inputs, in particular IT systems. Nevertheless, some subnational jurisdictions may be large enough to realize at least some of the advantages related to economies of scale; superior knowledge of local circumstances may also give an advantage to subnational administrations; more experimentation may take place etc. New developments in hardware and software have reduced the previous advantage of centralized information and processing systems. Unfortunately, so far, the available empirical evidence one way or the other is still very scarce.16 73. Taxpayer compliance costs generally can be reduced via centralized tax administration because of fewer offices to visit, less information to process, etc. However, decentralization may provide more proximity to subnational offices. In addition, a separate local tax administration regime can enhance the accountability of local officials to residents and taxpayers beyond the accountability that may exist when decentralized local taxes are collected by the central government but local government have an appropriate degree of policy discretion, in particular, control over tax rates. 74. From a political economy perspective, there is also an important question regarding incentives. If the central tax administration has the technical capability and cost efficiency, does it have the incentives to collect subnational taxes? Even if some degree of incentive compatibility may be introduced through appropriate contracting whereby the central tax administration gets monetarily compensated for the collection of subnational taxes, its effectiveness may be limited because central tax authorities may still prioritize their own collections and goals. Collections may also be affected by the presence of vertical tax externalities, whereby tax decisions by one level of government (for example, an increase in rates) may affect the size of the taxa bases and therefore collections of the other level of government. 75. Because of the different objectives that can be pursued, which can be weighted in different ways by decision makers, and because of the variety of political economy issues at play, we can expect a large variety of organizational models for tax administration in decentralized systems. The international experience in the vertical organization of tax administration shows a large variety of models and practices: • Single centralized tax authority enforcing all national and subnational taxes; this is the experience of Scandinavian countries and other countries like Russia and Belgium, and also common in more centralized unitary countries; • Independent tax authorities at different levels of government, with varying degrees of cooperation; this is common in large federal countries, including, Brazil and United States; • fully decentralized tax authorities, with all taxes—national and subnational—collected at the Subnational level; in practice this is the rarest with examples in Germany and Lao PDR, and historically in the Soviet Union and China before 1994; and • Mixed models of tax administration featuring variations of centralized and decentralized characteristics; this type can be found in Canada, Spain, or Switzerland. 16Overall, we need to be careful to assume that central tax administrations are always able to operate more efficiently. For example, Ebel and Taliercio (2005) report subnational tax administrations in East Asia that operate quite efficiently. 39 76. Which model that fits better is likely to depend on the specific tax assignment in a country. In summary, it would generally seem to be undesirable to decentralize the administration of taxes shared between the federal (central) and local governments or subnational taxes that piggyback on national taxes. In the case of exclusively subnational taxes, their administration can be assigned to the central tax administration, but this needs to be accompanied by incentive-compatible contracts. Subnational tax administrations are more desirable for exclusive subnational taxes for which there are information and enforcement advantages at the local level, such as, often argued, with the property tax. When the system is already working with multiple levels of tax administration, a high degree of dialog and cooperation between different levels of tax administration should be institutionalized. Box 6. Criteria for selecting tax revenues at the regional level Musgrave’s (1959) “three-roles” for government activities-- macroeconomic stabilization, redistribution of income, and resource allocation –are generally used to guide the assignment of revenue sources across different government levels. Tax instruments that can be used for economic stabilization and income distribution (under the “ability to pay” principle) are best assigned to the central government, while some of those related to allocative efficiency (how to best use the resources available to provide goods and services) under the “benefit principle” may be better assigned to regional governments. In general, available subnational taxes should not be excluded because of their lack of progressivity; redistribution should be a central government function. There are also some characteristics that are desirable in all taxes desirable, whether central or subnational These include: i. Revenue buoyancy, or revenues should change roughly in proportion to the economic base; ii. Fair or equitable, taxpayers in similar circumstances should be treated similarly and if not, according to their ‘ability to pay’; iii. Efficiency, relatively low administration and compliance costs and create a minimum of distortion in the economy; and iv. Political acceptance, be sensitive to the historical and institutional framework in a country. In addition, there are several other principles that are desirable for taxes assigned at the subnational level: First, the application of the “benefit principle” to the largest extent possible, relating revenue sources to the benefits from services provided. Second, tax bases should be relatively evenly distributed across jurisdictions, which helps to reduce fiscal disparities and the burden put on equalization grants. Third, relatively immobile tax bases, so to minimize tax competition among jurisdictions in a “race to the bottom. Fourth, geographically neutrality, so not to interfere with commerce or location of economic activity and also that local taxes are not “exported” (or primarily borne by residents in other jurisdictions). Fifth, subnational administrative feasibility. Sixth, subnational taxes should exhibit generally stable tax bases and revenue yields over the business cycle; revenue sources that are highly sensitive to general economic conditions (e.g. profit taxes) should be assigned to the central government, which has greater ability to deal with cyclical fluctuations in revenues through borrowing and other means. And, seventh, sub-national taxes should be highly visible so that tax burdens are clearly perceived by local residents. Source: Own elaboration 40 1.1.1.2. Typical sources of own revenues for regional governments 77. There are no many taxes that comply with all desirable criteria listed above. A compromise across criteria is generally needed. Nonetheless, even though we cannot select one single best tax assignment, the criteria allow us to select among better and worse tax assignment choices. Subnational governments are mostly prescribed to engage in activities ensuring a more efficient allocation of public resources, and therefore they should be assigned revenue sources for which it is easier to establish a link with the benefits received by residents from local government spending, that is following the “benefit principle.” Based on those principles, it is possible to draft a list of taxes that would make good choices for revenue assignments at the subnational level. The list of better choices of subnational taxes would include the following: 78. Fees and user charges: A straightforward way to raise revenue in accordance with the benefit principle is by charging user fees to cover the cost of providing specific local government services. Reliance on user fees tend be low in practice as they are potentially regressive. However, this is generally an equivocal argument. 79. Property taxes: 17 There is ample consensus that the property tax is among the best mainstays at the sub-national level, because it complies with most of the characteristics listed above. Also, local governments have a comparative advantage in identifying and valuing properties because they are familiar with the housing and land available there. But precisely for that reason, revenues from the property tax tend to be assigned to local governments as opposed to regional governments. 80. Vehicle and transportation taxes: These are generally good subnational taxes because, besides of the double dividend of combating pollution and traffic congestion, there is a strong link between the ownership of vehicles and the use of local and regional services and infrastructure (particularly roads). 81. Natural resource taxes (when resources are evenly distributed): Natural resource taxes can be justified at the regional and local level to the extent that extraction activities use local infrastructure (e.g., roads needed to transport heavy machinery and mined resources), place stress on other local infrastructure (temporary worker camps, hospital facilities required to treat injuries incurred by those working in this industry, and so on), and –depending on the type of extraction- may pollute the environment or cause other negative externalities increasing health costs of local residents. However uneven geographical distribution of resources can lead to fiscal disparities, and inefficient population migration and location of business. The volatility of revenues, reflecting world commodity prices, is also a negative factor. Natural resource taxes are usually levied by the central government and then (partially) shared with subnational governments according to devolution or derivation principles. 82. Local business taxes: Certain forms of business taxes, including business license fees, are justified at the regional and local levels as an indirect but administratively easier way to tax income of business owners (especially non-wage incomes), and as a benefit tax for the services and infrastructure provided by sub-national governments. However, to avoid economic distortions, any broad-based levy should be neutral to the factor mix, applying equally to labor (payroll) and capital (assets) used by businesses. 18 17 In practice there are several forms of the property tax. For example, some countries separate the taxation of land and improvements, or structures, and a few others tax only land values or rents. Although a tax on land tends to be more efficient, it also has less revenue potential and it is generally more difficult to administer properly, for example in terms of valuation or assessment of properties. Other types of property tax are “betterment levies” and other forma of land value -increment capture, very helpful in financing infrastructure. Also property transfer taxes can be relatively high revenue yielding but has the negative features of locking-in sales, inducing underreporting of sales prices, and high volatility in revenue flows. 18 An example is the so called a business value tax (BVT) in practice was Italy’s regional business tax (known as the IRAP) prio r to the elimination of payroll from the tax base in 2003. 41 83. Excises Taxes: Subject to the area size, cross-border trade and smuggling limitations, excise taxes have potential as piggyback taxes or special taxes at the regional level. Excises tend to be more politically acceptable, can be easily administered in coordination with national wholesalers as withholding agents, and allow for rates differentiated by region. For example, some OECD countries allow sub-national government surcharges on excises.19 Moreover, the benefit principle accords well with the assignment of (destination- based) excises on alcohol and tobacco to the sub-national level (to the extent that the latter is responsible for health care) and on vehicles and fuel (to the extent of sub-national government involvement in road construction and maintenance). The ability to charge differential rates across sub-national jurisdictions is of course limited by the possibility of cross-border trade and smuggling. The extent to which excise piggyback surtaxes can be used at the local level depends also on the technology of product distribution and points of sale. An interesting aspect of excise taxation at the sub-national level is the taxation of public utility services. There is significant revenue potential in some of these services, as in the case of electricity and phone services. 84. Flat-rate piggyback income taxes: Progressive income taxes are best assigned at the central government level for the reasons already mentioned. Nevertheless, the most commonly used form of regional income taxation internationally is a flat-rate income tax as a surtax or “piggyback” tax on the central government personal income tax. This type of tax is almost always collected by the central government administration and shared on a derivation basis.20 To enhance revenue autonomy, regional governments are allowed discretion in setting the flat rate between minimum and maximum rates, which are centrally legislated.21 Those tax instruments above are all good possibilities for regional tax autonomy. But there are also poor choices. 85. The list of considerably worse choices of subnational taxes includes corporate income taxes (CIT) and the value added tax (VAT).22 The CIT is ill-equipped for application at the regional level because its base—profits—is highly volatile and because there is no clean way to assign the tax on a derivation basis, since profits originate in the entire national territory and beyond. In the case of the VAT, it is also impossible to fairly allocate it on a derivation basis: the debiting and crediting of the VAT is likely to take place in different subnational jurisdictions.23 1.1.1.3. A practical guide for selecting subnational taxes 19 For example, in the Netherlands, provinces impose a surcharge on the motor vehicle tax levied by the central government. Provinces are free to set the rate of the surcharge, subject to a ceiling imposed by the central government. 20 Generally speaking, a regional income tax should be levied at the place of residence because it is there where most taxpayers consume sub-national government services. However, because of administrative convenience, sub-national piggyback taxes are often withheld at source at the place of work by employees. However, it is quite feasible to distribute the funds according to where workers reside. 21 Other forms of tax autonomy are practiced, such as the ability to modify the base of the tax by providing more or less deductions, exemptions and so on. 22 Basically, only three large federal countries have introduced sub-national VATs: Brazil, Canada and India. The mixed experience from these countries has served for many years as an example of the difficulties facing any other country contemplating the introduction of a sub-national VAT. 23 Note, however, it is perfectly feasible to share VAT revenues with sub-national units using a formula; for example, the VAT can be shared on the basis of population (as in Germany and Belarus), or on the basis of the regional shares in aggregate consumption (as in Canada’s Maritime Provinces, Japan or Spain). Even though the CIT could also be shared using a formula, the attraction to do is severely diminished by the instability (over the business cycle) of its yield. Revenue sharing is discussed further below in the section on transfers. 42 86. The practice of tax assignments varies widely across countries with many different taxes being used. Given the different criteria that subnational taxes should meet, the choice of particular taxes comes down to a comparison of the advantages and disadvantages that each of those taxes offer. In fact, it is possible to give a score to most of the taxes commonly used at the subnational level. This is done in Table 1 where three letters are used for high, medium and low potential of each tax in supporting a set of desirable characteristics, including revenue potential, ability to fit the benefit principle, non-exportability, and so on. With some caveats the scores presented in Table 1 can work as a guideline for policy makers interested in identifying desirable traits in a long list of potential taxes that are used in the international practice to provide large cities with tax autonomy. Those taxes that have dubbed as “good choices” are quite commonly used in many decentralized countries and they expectedly tend to score higher with more of the desirable properties, while those dubbed as “potentially bad choices are less commonly used and they tend to score lower with many of the desirable properties. The category of “miscellaneous” refers to decentralized taxes that are less frequently used but that depending on the circumstances they can provide some revenues while achieving some desirable properties. Among the desirable properties on which we would like decentralized taxes to score high, the table includes: good revenue potential, with high elasticity or buoyancy (meaning that revenues will grow proportionally with the economy), high adaptability to the benefit principle (so that those paying the tax are those using the public services), an even geographical distribution of tax bases (so that there are less disparities in per capita revenues across jurisdictions), fairness and equity in the distribution of tax burdens (those with higher incomes pay proportionally more), lower costs of administration and compliance by taxpayers, higher acceptance and better perception about the tax among officials and the public, and high visibility of the taxes so that can facilitates greater accountability by public officials. Among those properties that are negative, less of which is desirable, we have inter-jurisdictional mobility of the tax bases, negative efficiency effects in the allocation of economic resources, high sensitivity to fluctuations with the business cycle in the economy, the risk of incentivizing or facilitating corrupt practices among tax administrators and taxpayers, and the risk that tax burdens may be exported to other taxpayers in other jurisdictions (so bringing less accountability and fiscal responsibility). Obviously, every tax offers a different package of properties. The art of tax assignment is to select a basket of taxes that can offer a desirable package of properties. Also, quite obviously, not all properties will be weighted the same. Revenue yield, stability and elasticity may be deemed more important than acceptability and administration costs, for example. 43 Table 1. Advantages and Disadvantages of Observed Subnational Taxes24 Verti Cost of Acceptability Poten Sensi Adapta Even Buoyancy cal Administ Potent Visibility/ Mobility tial tivity bility to distrib Compli Expor Revenue Elasticity equit ration ial for By Local of Tax Effici to benefit ution of ance tabilit Potential y (By corrup By Privat Accounta Potential Base ency Cycl principl tax Costs y /fair subnation tion Politic e bility costs e e base Tax ness al govt.) ians Sector Good Choices Real Estate Property Tax L/M L L L L H M M M/H L M L L L H and Betterment Levies Property Transfer Tax M/H H L M/H H M M M L/M L M M/H M L H Business Tax M M/H M/H H L M M L/M M L/M M/H Individual Income and H H L/M M/H25 M/H H L H H M M/H H L L26 H Payroll Taxes Sales Tax H H L L/M M/H M/H H L M M/H M/H H M M M/H Vehicle Tax L/M M L/M L L M M/H H M M L M/H M L H Construction Tax M M/H L M/H H M/H M M/H M M M/H H L/M L H Transportation tax L L/M L M L/M M/H L M H M/H M M L L M Excise Taxes Alcohol Tax L M/H L L M M H L/M L L M H M/H M M/H Electricity tax L/M H L L/M M H M/H L/M L L L H L L H General Excise Tax L/M M/H M/H M/H M M/H L L/M M L L L/M M/H M M Gasoline M H M/H M M/H H M/H H L L L H L/M L/M H 24 The table is based on the theoretical principles of tax assignments, the experience of numerous countries, and also numerous papers in the public finance literature that have examined the properties and performance of the different taxes. 25 In the case of payroll (wage) taxes the distortion effects will tend to be higher. 26 If levied on the place of work, it may be exported to nonresidents. This is not inappropriate, if the tax reflects services provided to non-residents, e.g., commuters. Green Tax L M L/M L M M/H L/M M M/H M L/M H H M H Telecommunications tax L/M H L M M M/H M M/H L L L H M L H Possibly Bad Choices Corporate Income Tax M/H H H H H L L M/H H M/H M/H H L/M H L VAT H H M M H M/H L L/M M/H L/M L/M L/M L/M L/M M Miscellaneous Advertisement Tax L M L L M M L M M/H M M/H H M/H M M Amusement tax L M M M H M/H L M L M M H L M H Financial Tax H M H H L L L L H L L Gambling Tax L H H L M L/M L M/H L L H H M/H M/H M Hotel tax L H M M H H L M L M L/M H L H L Insurance tax L M L M L M L M L L L H L/M L M Natural resource tax L H L M M M/H L M M L M H L H H Stamp tax L/M M H M L L L L H L L Inheritance /wealth Tax L L M/H L L L M H M M M/H L L L H Note: H=High potential, M=medium potential, L=Low potential Source: Martinez-Vazquez (2016). 45 5.2.2. Transfers and grants27 Transfers and grants are important components of the revenue architecture for subnational governments. They are necessary to help close vertical and horizontal imbalances stemming from either inadequate or unequal own resources (through revenue sharing and equalization transfers)28, finance certain expenditure programs on behalf of the central government and specific investment programs. 87. There are different types and forms of transfers and grants. The most relevant ones are described below. 5.2.2.1. Revenue sharing 88. Revenue sharing plays an important role in most decentralized fiscal systems. Revenue sharing is mostly implemented on a derivation basis to pursue devolution objectives, and mostly used to help with the closing of vertical fiscal imbalances (See Box 7). 89. Most OECD countries, but also most developing decentralized countries, implement some form of revenue sharing. Countries that implement revenue sharing–allowing regional governments to keep some of the funds where they were generated–intend to provide incentives to develop the regional economies and also provide some sense of national balance and cohesion. For example, Spain provides for revenue sharing in central VAT revenues by allocating them according to the share of aggregate consumption in regional GDP. That allocation rule clearly pursues a devolution objective. This same country also shares personal income tax revenues on a derivation basis by allowing regional government to retain a share of the receipts according to the place of residence of taxpayers. This latter again equally pursues a devolution objective. In another example, Germany implements VAT revenue sharing with the states according to their share of population. This latter distribution criterion tends to mix devolution and distribution objectives. 90. Revenue sharing following devolution objectives clearly benefits the most economically dynamic regions of the country. And it is generally inevitable that only some parts of any country will tend to be the engines of growth. That is also likely to true for Chile. Revenue sharing on a derivation basis allows a country to recognize the role of the most economically dynamic regions and provides them with a salient incentive to continue to play that role. But the larger inequality or horizontal fiscal imbalances so created can be addressed—as is generally the case in the international practice—with a separate instrument, putting in place a strong equalization grant. Well-designed equalization grants recognize automatically in their formula revenue sharing funds as part of the fiscal capacity of the regional governments, which reduces, if not eliminates, the equalization funds they eventually receive. 91. Tax sharing on a derivation basis imposes restrictions on the type of central tax revenue sources that can be used for that purpose. In most countries it would be extremely hard to do a transparent allocation of corporate income taxes on a derivation basis given that most of that tax is paid at the location of headquarters, but the actual economic activity happens elsewhere in the country where production factories are located and sales actually occur. In addition, the corporate income tax is not a desirable tax to share 27The discussion in this section is partially based on Martinez-Vazquez (2019). 28While both revenue sharing and equalization grants help close the vertical imbalance, only equalization grants contribute to close horizontal imbalances. Note, however, that many Latin American countries, including large federal countries like Argentina and Brazil, and also other countries around the world, like India, use a single instrument to pursue both objectives. The result is that in the end it is not clear what is being achieved in any particular dimension or objective. with regional governments because its revenues are generally quite unstable over the business cycle and that would put pressure on regional budgets, which the central government is better equipped to withstand. 92. There are other types of taxes that can be more easily shared on a derivation basis with the regional governments, such as in the case of the personal income tax. The personal income tax and the VAT are the largest most commonly shared taxes. Note that in the case of the VAT, it is also generally impossible to allocate it to the regional governments on a pure derivation basis, because the tax is debited and credited in many different parts of the country and it is hard to establish where the value added is generated. Also, as in the case of the corporate income tax, it is likely that most of the tax would be allocated at the place of headquarters of companies, when the tax is generated across the entire country. But it is easy to approximate origin by utilizing a formula using the share of regional consumption in GDP. Of course, a similar formula could be used to allocate the corporate income tax (for example the regional share of payroll or final sales). But, this latter is not done in the international practice because, as mentioned, the corporate income tax is subject to drastic oscillations over the business cycles, and subjecting subnational governments to sharp revenue swims is generally undesirable. Box 7. Vertical fiscal imbalances Conventionally, vertical imbalances have been measured in two different ways. The first is to look at the surplus or deficit position of each consolidated level of government, before borrowing but after all revenue sharing and transfers to subnational governments has been implemented. This measure tends to be biased towards central or federal government mismatch (too little funding goes to the central government vis-à-vis sub-national governments) because in many countries, either by law or by practice, sub-national governments operate more conservatively than the national government. The central government can borrow more freely to cover current operations. In contrast, local governments in most countries are only allowed to borrow for capital investment purposes or not at all. A second way to measure fiscal imbalance is to examine what share of sub-national government expenditures is financed with sources of revenues under the control of sub-national governments. The coefficient for vertical imbalance is calculated as one minus the share of the sub-national government expenditures that are financed from sources of revenues not controlled by the sub-national government. Typically, there is disagreement on how exactly to define those revenue sources that are controlled by a higher level of government and therefore different versions of the coefficient are computed. By any of these measures it is often the case that there is considerable vertical imbalance because of the low revenue autonomy at the sub-national level. Source: Own elaboration 5.2.2.2. Equalization transfers 93. Because the level of economic activity and wealth of regions in countries tend show marked differences, the assignment of own revenues (good in their own merit) tends to produce horizontal fiscal imbalances (Box 8). The same is true with the introduction of revenue sharing on a derivation basis (also good in its own merit). The presence of these horizontal fiscal disparities, if not addressed, is undesirable on both equity grounds because citizens in different locations would have different access to basic services, and also on efficiency grounds, since those differences in access to public services would lead to purely undesirable fiscal migration (as opposed to economic migrations for job opportunities, etc.). It is for these reasons that most countries around the world have introduce equalization grant systems at the regional level. Box 8. Horizontal fiscal imbalances 47 There are different ways to measure horizontal fiscal disparities. The most immediate approach is simply to compare revenues and expenditures per capita across jurisdictions. However, both actual per capita revenues and expenditures may be misleading indicators of true fiscal disparities. For example, revenues per capita may be higher in a relatively poorer region or locality because of a greater effort exercised in enforcing and collecting taxes. Similarly, comparisons of actual expenditures per capita may be misleading. For example, expenditures per capita may be higher in a particular region or locality because of the higher expenditure needs in those jurisdictions as represented by vulnerable groups such as under- age or retired populations, the unemployed or simply higher costs of provision because of weather, prices, or communications. Thus, a more careful comparison of horizontal fiscal imbalances calls for, on the revenue side, the use of measures of fiscal capacity or the size of potential tax bases in per capita terms, rather than simply per capita revenue collections. On the expenditure side, the measure of horizontal fiscal imbalances calls for the adjustment of actual expenditures for a variety of measures for expenditure needs. The measurement of horizontal fiscal imbalances or disparities based on fiscal capacity and expenditure needs, as opposed to actual revenues and expenditures, is often complicated by the lack of data availability. In practice, a useful approach to the measurement of horizontal fiscal imbalances is to use a combination of comparisons of actual revenues and expenditures per capita and different proxies for the measurement of fiscal capacity (such as income per capita or regional gross domestic product per capita) and expenditure needs (such as the relative presence of population groups in more need or differences in cost of provision of standard services). By how much should horizontal fiscal disparities be reduced? This is a difficult question which ultimately has to be answered in political terms. The degree of equalization has to reflect a national consensus made through representative institutions on how much to take from relatively richer regions and from national or federal government budgets to support worse-off regions. It is an issue of national solidarity and trade- off of possible faster economic growth (which may occur by leaving more resources in better off regions) with greater equality or equity across the national territory. But in addition, reducing horizontal disparities is not only important from an equity stand. The disparity in the level of public services across jurisdictions will also cause inefficiencies by inducing not always desirable migration of populations. Source: Own elaboration Objectives of fiscal equalization transfers: 94. The first step is to have a clear vision of the objective pursued with an equalization grant. The most commonly accepted objective is to allow regional governments to provide to their residents a similar level of access to a standard package of public services when they exercise an average level of tax effort. The objective of equalization grants is not to equalize income per capita or the level of economic development across regions. Believing the latter is actually a commonly made mistake; but economic development and per capita income levels are much more complex outcomes, about which equalization grants can only help in limited ways. Basic architecture: 95. Two fundamental steps go into the designing the basic architecture of an equalization grant. The first is to determine the size of the of the equalization pool of funds. The second the distribution or allocation formula. 96. The pool of equalization funds: The two main approaches to setting up the equalization pool of funds is to do it ad hoc every year in the national budget or to use a fix rule or formula for a more stable and certain determination of the pool of funds. There are countries that use an ad hoc approach but there are certainly examples of other countries in the international practice that fix the size of the equalization 48 pool in the law. For example, Australia exclusively assigns to the pool of equalization transfers 100 percent of the collected VAT revenues. Other countries that use a defined sharing rate of the revenue pool include Canada, Germany, Switzerland, Russia and most of the Nordic European Countries. There are two main types of arguments in favor of this policy reform. The first is an argument about the increased stability and predictability of the transferred revenues that would accrue to the states. The second is an argument about incentives, to eliminate an implicit soft budget constraint faced by the states —they can always fight for a higher revenue share and do less about increasing their own tax efforts. This would appear to have been an overlooked important issue in the common discussion of this issue. 97. There are, on the other hand, also arguments for the need for even more flexible terms in determining the pool of funds or the sharing rate of central revenues with the states. These arguments are mainly centered on preserving the degrees of freedom for the central government to implement macroeconomic fiscal stabilization policies. Over time, the type of argument has lost strength and validity, as central governments have been increasingly seen as the ablest to absorb cyclical risks and implement offsetting policies, including borrowing, which should be much more limited to subnational governments. 98. Allocation formula: Regarding the distribution formula, the state of the art in the design of equalization transfers in the international practice is the “fiscal gap” approach, defined as the difference between separate estimates of expenditure needs and fiscal capacity of the states. An increasing number of countries have adopted this methodology. Among developed OECD countries we find Australia, Canada for the Northern Territories, Italy, Japan, Korea, United Kingdom, and in many U.S. states; among countries in transitions, China, Latvia, Russia, Ukraine and Vietnam; and among developing countries, Indonesia, Peru, and Uganda. Perhaps the state of the art in the design of a fiscal gap methodology is Australia, where the fiscal gap equalization formula is administered by an autonomous commission, and revenue capacity and expenditure needs are meticulously calculated every year for a long list of revenues and expenditure areas. However, other similar approaches are also used. For example, Canada implements equalization for the its provinces (as opposed to the territories) only on the basis of fiscal capacity per capita. Germany, and other countries such as Poland and Spain use yet another variation of the methodology by equalizing fiscal capacity per adjusted population (instead of simply per capita), where adjustments to the actual population are made to reflect differences in expenditure needs. In Brazil, like in Argentina and other Latin American countries, partial equalization is implemented by using a weighted index formula approach containing variables that proxy needs and capacity. The main design features of the fiscal gap approach: Quantifying expenditure needs and fiscal capacity and the distribution of available funds 99. There are different methodologies that are used to compute expenditure needs and fiscal capacity across subnational government, as there are also different ways to apply their difference—the fiscal gap— to implementing the actual distribution of the available equalization funds among subnational governments. Those techniques and methods are described in some detail in the Appendix. 100. Several methodologies are used to estimate tax capacity across subnational government, and the preferred more sophisticated ones do a good job at estimating the potential revenues that can be obtained from the tax bases assigned to the sub-national government (when they exercise an average or maximum level of collection effort). This is important because there is a need to avoid perverse incentives encouraging lower actual collections. The actual measurement of fiscal capacity includes those potential tax revenues plus revenue sharing and possibly other unconditional transfers. For these latter sources of funds actual quantities can be used since central governments control their amounts. The most credible methodologies to estimate tax capacity include: (i) the Representative Revenue System (RRS), utilized in in Canada and in the U.S., which uses information on tax bases and average collection effort; (ii) the stochastic frontier estimation of potential maximum revenues, which assumes subnational governments deviate from the 49 optimal collection levels because of lower administration collection efforts; and (iii) using basic proxies for the local ability to tax, such as for example the per capita level of personal income. Other simpler methods, such as lagged own revenue collections or average of past collection ratios, are less reliable and introduce incentive problems. 101. Regarding expenditure needs, the exercise is to quantify the funding that would be necessary to cover all expenditure responsibilities assigned to the sub-national government at a common standard level of service provision. Typically, the estimation of expenditure needs is restricted to those arising from current expenditure needs—therefore excluding those arising from capital expenditure needs. The reason is that capital needs tend to be more complex, lumpy and discontinuous, and generally they are better addressed separately via the utilization of capital investment grants. 102. The international practice shows also in this case several methodologies that can be followed. Among the most preferred ones we find the “per client (top-down) financial expenditure norms”, which specify per client expenditure standards, either from a normative viewpoint or form a historical and affordability viewpoint (for example, derived by dividing aggregate level of expenditure across all subnational governments in each functional area by the number of clients or users of that function at a national level). Equal in the level of sophistication is the “bottom-up costing of baskets of standardized inputs” methodology, which costs standardized baskets of state government services by functional area. A third desirable methodology is the “regression-based representative expenditure system (RES)”. It is a bit more complex and data intensive than the previous two and it uses regression analysis using data on expenditure per function and drivers of those expenditures and the results can be interpreted as the amount of money that a subnational government would have spent in a particular service function in order to provide a standard level of service. Less accurate methodologies utilized in the quantification of expenditure needs include: “weighted indexes of expenditure need proxies”, “lagged expenditure values”, which rely on information on what has been spent in the recent past for each function; and “equal per capita expenditure norms”, which would the easiest to compute but that generally cannot be justified. 103. As already mentioned above, once the expenditure needs and fiscal capacity have been estimated, the “fiscal gap” or the difference between the two can be computed for each subnational government. The pool of funds available for equalization can be distributed among those jurisdictions –proportionally to the fiscal gap or in some other manner—for which expenditure needs exceed fiscal capacity. 104. One final issue concerns the standard of equalization. This is fundamentally a politically decision constrained by the pool of resources dedicated to equalization. Box 9. Adjusted population index As noted above, within the fiscal gap methodology, there are some countries which only equalize differences in fiscal capacity per capita and ignore potential differences in expenditure needs, as for example is the case of Canada. On the other hand, there are several countries that go a step beyond and use a fiscal capacity per adjusted population approach, as in the cases of Germany, Poland and Spain. This methodology requires a set of calculations. The first is the estimation of fiscal capacity for which the different approaches discussed under the previous issue could be utilized. The second is to estimate the “adjusted” (by expenditure needs) population of each state. Thus this step does involve some sort of estimation of expenditure needs. However, the estimation of expenditure needs in this case does not need to be as involved or accurate as was described above for the fiscal gap methodology. The basic intuition of the fiscal capacity per adjusted population approach is that the adjusted population in the state will exceed the actual population of that state when its expenditure needs are above the 50 average expenditure needs of all the states, while the adjusted population would be smaller when the state’s expenditure needs are below the average. Those expenditure needs may be measured via a weighted index, comprising a number of variables that approximate expenditure requirements for the various functions of the states and relative weights that reflect the relative importance of those functions in the states’ consolidated budgets. The adjusted population for each state can be obtained by multiplying the value of the weighted index for the state by the total population of the sates –that is, the national population. The ratio of adjusted population to actual population in the state denotes whether fiscal capacity per adjusted population will go up or down and therefore whether the state will be eligible for more or less equalization funds. The next step is to rank all states according to their fiscal capacity by adjusted population. The final step is how to use this ranking series to distribute the available funds for equalization. There are many ways to do this. For example, funds could be distributed proportionally to the differences in the ratio of fiscal capacity per adjusted population for the states with ratios below the average, or according to the distance to the state with the highest ratio, or according to a max-min criterion with the funds utilized to bring up the lowest ranking states. Source: Own elaboration 5.2.2.3. Conditional grants 105. There are different types of conditional grants; block grants and specific grants, which can be matching and not; and conditionalities can be ex ante and ex post –like in the case of performance-based grants. 106. Conditional grants design should always start with the fundamental question of what type of instrument (that is, conditional grant) would be best fit to achieve the goal or purpose that is being sought in the first place. From this perspective, it is important to recognize that there is no absolute dominance of one type of conditional grants over others, for example, of block grants over specific grants, or performance- based grants over ex ante conditional grants. Specific purpose grants (with ex ante conditionality or ex post performance-oriented conditionality) may be what is needed depending on the objectives of the government. 107. It is important to realize that one main drawback with all types of conditional grants is that they tend to limit the autonomy of subnational governments. If decentralization offers the advantage of achieving a more efficient allocation of resources by letting subnational government exercise their discretion and autonomy in their budget decisions, obviously conditional grants–in contrast to own revenues and unconditional grants–tend to limit that autonomy and therefore may be considered less desirable. 108. On the other hand, there are legitimate reasons for central governments to introduce conditional grants. In the first place, conditional grants can be used to finance delegated functions to subnational governments. Delegated functions mean that the central government is still responsible for those services but chooses to use the services of subnational government as agents to deliver that central government responsibility. These are known as “pass through” or “agency” or “cost reimbursement” grants. Central governments may also pursue the objective of guaranteeing the minimum level provision of what can be considered “national merit goods”, for example some levels of education or health care. In this sense central governments may use specific grants–for example, for school lunches or vaccination programs—which require that subnational governments spend that money in those specific areas, or else the central government can use “block grants” funding additional subnational expenditures generally in education or health, etc. Of course, block grants give more flexibility to subnational government in deciding how to 51 spend the funds. That is desirable from the viewpoint of autonomy. However, the narrower goal of the central government, that the funds are specifically spent in a certain manner, may well override the general desire for more subnational autonomy. That is also the case when central governments try to address externalities (negative and positive) across subnational governments. In this latter sense a specific grant would be justified if, for example, central authorities try to stimulate spending on a sewage treatment plant for subnational governments that otherwise may pollute upstream river waters. 109. Internationally over several decades, however, central governments in many countries overused specific grants to address a multiplicity of issues, in ways that were not always warranted. As a reaction to that, over the past several decades there has been an international trend towards the simplification of transfers systems in order to provide subnational governments with more autonomy in the use of transferred funds (Box 10). In fact, this general trend can be decomposed into sub-trends both working in the same direction of providing greater subnational authority: first an increase in the share of unconditional grants over conditional grants, and second, within conditional grants, an increase in general purpose block grants with a decrease in specific purpose conditional grants. For several decades before, the experience of many countries had been to see transfer systems grow into a jungle-like mix of specific purpose grants, many of them too small, costly—at time not possible-- to administer, imposing high cost compliance and reporting costs on subnational governments, and often times overlapping in targeted roles with contradictory objectives. Box 10. International experience with streamlining of transfer system Switzerland: Moving away from earmarked grants: The country had a system in which equalization between the different cantons was mostly achieved via earmarked grants. This meant that specific purpose financing and equalization were often strictly tied together. In the new grant system, implemented in 2004, financing, subsidization and equalization are separated, and earmarked matching grants have largely been abolished and earmarked non-matching grants were sharply reduced--from more than 50 to 17 and a number of perverse incentives have been removed from the grant system. Simplification reforms were also implemented in the Netherlands, Sweden and the United Kingdom with the conversion of earmarked grants into block general grants. The hope was that sub-national governments would continue the existing spending pattern after the change from earmarked to block grants. This has been the case in Sweden, for example, with the financing of child care. However, in the United Kingdom and the Netherlands, the central government entered in agreements subnational governments to target the yearly increase of the non-earmarked grant for central government priorities: education policy in the United Kingdom; health, education and police in the Netherlands. Source: Bergvall et al. (2006) 110. Importantly, and as briefly mentioned above, beyond seeking greater autonomy for decision making among subnational governments, other reasons have included avoiding the over burdening of the administrative capacity of subnational governments and implicitly assuming responsibility for the competencies of those subnational governments. Central governments also have sought to benefit from establishing a clearer set of national priorities in the use of funds, easing monitoring costs, and more easily gauging progress in the achievement of their goals. In addition, earmarked grants have been identified in some cases to introduce perverse spending incentives among subnational governments (Box 11). Box 11. Specific Earmarked Transfers Can Lead to significant Distortions 52 In Austria the central government gave earmarked transfers to the states for teachers’ salaries, with the states being responsible for recruiting teachers and negotiating teachers’ salaries, but the central government being responsible for funding the salaries. This created two inefficient incentives for the states: to employ too many teachers and to negotiate salaries that are too high. Recognizing the problem, a national limit on the student-teacher ratio was introduced in 2000. Mexico has had a similar problem with respect to education transfers. The federal government provides the states with transfers for education, based primarily on the number of teachers they employ. One difference with Austria is that salaries are negotiated through a stronger national teachers’ union. The social assistance law was introduced in the Netherlands in 2004 to create incentives for municipalities to reduce the number of people on social assistance. Before this reform the central government largely reimbursed municipalities for the social assistance-benefits they paid. This procedure gave municipalities little incentive to constrain expenditures. Since the reform was introduced, municipalities have borne the full responsibility for social assistance and are reimbursed for the cost through a block grant. The level of the block grant is decided by macroeconomic variables that municipalities cannot control (if the general economic situation worsens it increases and vice versa). The grant is no longer based on actual expenditures but, depending on the size of the municipality, on either a set of relevant criteria (large municipalities) or on historical cost data (small municipalities). Source: Bergvall et al. (2006) 111. The international experience offers a rich body of practice that can inform the choices of conditional grant system that a country wants to adopt. The first thing to note is that, given the many forms to fund transfers, the types of distribution approaches and the types of control that can be imposed, there may be over 60 types of possible transfers. Regarding the types of restrictions in the use of funds distributed, for example, there are: (i) unconditional use of funds; (ii) block allocation allowing for discretion in the use of funds but only within a general sector or identified area;29 (iii) highly restrictive ex-ante conditions, as in the case of specific use grants, with detailed use restrictions; and (iv) performance based grants with ex- post conditionality. 112. One of the essences of conditional transfers is that more often than not they become a source of conflict between central and subnational governments because of the different priorities given to particular expenditures. The resolution of that conflict often entails the use of conditional matching grants, which allows central governments to impose their priorities while subnational governments are free to exercise their own budget decisions, but they tend to comply because of the effectively lower cost of the services preferred by the central authorities. 113. Conditional grants—whether block or specific—typically are best implemented (when feasible) on a “capitation basis” (i.e., per student and per inhabitant), with this basis appropriately modified for costs or needs differences.30 There are many different types and intensities of conditionalities (see Box 12). Using fewer conditions and making those transfers block grants generally has the added advantage, as pointed out above, of allowing subnational authorities to exercise budgetary autonomy in terms of setting some spending priorities and selecting the most efficient method of service delivery. 29 Internationally, there is some terminology confusion regarding block grants because different countries use the same terminology for quite different types of grants. As Bahl (2009) points out, for exampl e, in the United States, “block grants” means transfers that can be spent for wide spending categories such as health, education, or environmental protection. On the other hand, in practically all Nordic countries in Europe “block grants” are taken to mean unconditional or general-purpose transfers without any conditionality attached to where the funds should be used. The meaning followed here for block grant is one with loose general conditionality. 30 In particular, using existing capacity as the basis of the grants should be avoided since otherwise transfers would be adding the inequity of providing less operating funds to those jurisdictions with already more inadequate physical facilities. 53 Box 12. Types and effects of conditionalities Service delivery conditions are those that specify the function for which and how the funds are to be spent. They can range from being very specific to being very general. When conditional grants are simply given for a function with few or no specific conditions attached, they are referred to as block grants. Specific grants, on the other hand, carry detailed service delivery conditions, defining how services are to be delivered as well as what services are to be delivered. And the conditions can relate to the client group as a whole or to a specific subset of that group. There are several types of conditions. They may involve financial conditions, administrative conditions, informational conditions and public relations conditions. The most frequently used financial condition is a matching requirement—the funds to be provided by the recipient- which leverage the impact on subnational expenditure choices. Beyond the benefits of fund additionality (especially when a ‘maintenance of effort’ is required) and securing more subnational ownership, matching requirements can also have drawbacks. Subnational governments may have different abilities to co-finance; this may call for graduated (for fiscal capacity) matching rates. The “cheaper” costs associated with mat ching may also lead to less careful or efficient subnational spending. A frequent problem is that central agencies are often late in designing and arranging funding for their matching grant programs; this leads to hoarding of liquid budgetary resources and hurried spending at the end of the fiscal year. Administrative conditions are most frequently about the environment in which services are provided, the efficiency in production and service delivery, and conformance with procurement standards. They are clearly intended to increase the quality of the process and services provided but they are the most likely to be considered as an undesirable interference in subnational autonomy. On the other hand, information conditions relate to the data the recipient subnational government must provide to the central agencies regarding the use and impact effectiveness of the spent funds. Finally, public relations conditions, which typically have no bearing on service delivery efficiency, may be demanded by officials to take credit for the changes. Source: Searle and Martinez-Vazquez (2006). 114. However, again, this is not to say that block grants are always a superior instrument compared to specific purpose grants. In fact, there are legitimate central government objectives that can be significantly better pursued by means of specific purpose grants. This means, also again, that rather than following a blind rule toward conditional grant design, central governments can do better by carefully selecting the instrument which best fits the chosen objective. 115. In fact, in recent years, many countries have transitioned from long lists of earmarked grants to simplify much shorter portfolios of conditional block grants.31 However, this transition has not been permanent, and in many cases experiencing reversals. One of the reasons for the reversal to specific use grants from block grants is associated with the political economy of grant making. For example, Borge (2009) explains that reversal in Norway is what he calls the “blame game” between the central and subnational governments. The former got blamed for providing insufficient funding necessary for critical 31However, the strength of this trend may have been exaggerated. A relatively recent survey of the evolution of intergovernmental grants in Sweden, United Kingdom Italy, and the United States finds that that there is almost no common feature or pattern in the choice between general block grants and earmarked specific grants (Kim 2009 b).See also Blom-Hansen (2009) and Hermansson (2009) for accounts on how Nordic European countries implemented massive streamlining of specific grants into block grants in the 1980s but how the number and importance of specific grants came back in later years. Additionally, Blöchliger and Vammalle (2009) find in a survey on intergovernmental grants in OECD countries that on average the importance of earmarked and non- earmarked grants were quite similar in 2006. 54 services, while the latter got blamed for the wrong spending priorities and low tax effort. With subnational governments successfully gaming the system, a soft budget constraint can emerge. This leads central authorities to revert to using specific grants instead of block grants. 116. It is important to note that the adoption of block grants has not always been accompanied by the greater discretion and flexibility at the subnational level. The movement from specific grants to block grants may be accompanied by other policy changes. An important lesson is that the control of subnational government spending by central authorities implied in earmarked specific grants can be exercised by other means, like additional regulations. In fact, the movement toward block grants in some countries did not necessarily produce greater discretion by subnational authorities because the introduction of block grants was accompanied by an increase in regulations by the central governments.32 117. The transition from specific to block grants has also been accompanied by controversy about the pros and cons of conditional block grants versus specific earmarked grants. What we have learned is that there are both advantages and disadvantages to using both types of conditional grants and the major lesson learned is that there is no intrinsic superiority of one type of grant over the other in all relevant dimensions. Each grant needs to be designed with the features that can best achieve the results and objectives that policymakers have in mind. Successful results may require the flexibility of block conditional grants or, instead, they may require the narrow focus and control of specific grants. 118. The conclusion that there is no universal type of grant that is superior to all others is congruent with the realization that, in considering different types of grants, policymakers walk along a clear tradeoff line. At one end they have more (vertical) accountability and control (from specific earmarked grants) to increased efficiency in the final use of funds (in the case of conditional block grants) to greater effectiveness in results (in the case of performance-based grants). Where one selects to be along those tradeoffs is a matter of constraints and particular objectives. Administrative capacity at the subnational level, degree of institutional fiscal autonomy and the extent of horizontal accountability can differ considerably across countries. Block conditional grants generally presume the presence of administrative capacity, fiscal autonomy, and subnational horizontal accountability. Significant deficiencies in those areas may call for more reliance on specific grants. 119. So, where does the international best practice stand on the use of specific versus block grants? There are clear disadvantages to using specific grants: they tend to be costlier to administer, they are less respectful of subnational autonomy, they tend to be less predictable budget-wise, and are more subject to rent seeking and clientelistic threats. On the other hand, specific earmarked grants may work well in cases where the central government has incomplete information on costs and expenditure need differences. They can also contribute to intergovernmental cooperation and be more conducive to developing more horizontal accountability given the higher visibility of the expenditure-revenue link. Specific grants may also be less likely to lead to a soft budget problem ---as in the case of discretionary changes in block grants—and therefore lead to more effective cost containment. Finally, specific grants generally offer central governments more assurance that its preferences will be respected in the implementation of redistributive social programs (Smart and Bird 2009; Lotz 2009), and Kim 2009b). 5.2.2.4. Capital grants 120. The international practice is generally to use distinctive separate capital grants or transfers in support of subnational governments’ needs to build public capital infrastructure. Because of their 32 See, for example, Blom-Hansen (2009) for the case of Denmark and Brosio and Piperno (2009) for Italy. 55 “lumpiness” and non-recurrent nature, the needs for capital infrastructure cannot be adequately taken into account in the recurrent equalization transfer or with ordinary conditional grants. 121. The most recent advances in public budget management call for the integration of all expenditure categories in the same budget –that is not to have a separate capital budget –so that all budget priorities can be compared, and overall efficient decisions can be made on the allocation of scarce fiscal resources. However, when it comes to supporting subnational governments with their capital expenditure needs, it is generally necessary to count on separate instruments. In particular, the international practice is not to mix capital expenditure need support with the equalization transfers –which only consider expenditure needs arising for recurrent obligations and fiscal capacity from recurrent funding. Capital infrastructure investment needs are characterized by their lumpiness and discontinuous nature, features that are very different from those of recurrent expenditure needs. In addition, from the financing potential side, in the case of capital infrastructure it is generally necessary to consider the role that credit and borrowing can play. However, access to credit cannot be easily incorporated in the computation of the fiscal capacity measures typically incorporated in equalization grant formulas. And, in addition, states have very different degrees of access to outside capital and credit markets. 122. The international practice with capital transfers is highly varied (Box 4). In terms of objectives, capital transfers are generally designed to assist subnational governments with financing constraints for lumpy capital, ameliorate significantly different infrastructure endowments across those units, pursue and support central government specific sectoral objectives, and possibly to address externalities across subnational governments. Two major policy biases need to be openly addressed in the design of capital grants. First, the belief that capital expenditures are always more efficient than recurrent expenditures and second, a disregard for the lack of maintenance of existing subnational government infrastructure. In the latter regard, matching grant arrangements are generally used to help subnational governments to take ownership of capital infrastructure projects and also to get them more involved in properly maintain existing infrastructure.33 123. In terms of design, capital grants vary by the degree of flexibility in the use of the funds. They can either be specific project-based grants, which tend to be closely administered and monitored by line ministries, and categorical or block grants, which can be designed with strong equalization features and which give much more discretion to subnational governments for which projects the funds can be used. These two approaches can exist side by side depending on the objectives of the central government and the nature of the projects. 124. In terms of their distribution or the way funds are allocated, capital grants can be classified as those distributed by ad hoc decisions and negotiations, by using pre-established formulas, and by competition processes with defined application procedures.34 No single approach is best in all cases, but non-transparent, highly detailed, and discretionary procedures should be avoided to prevent inefficient allocations and corrupt patterns. Formulas based on needs and client base are often quite feasible and commonly used. In Australia, for example, funding for school buildings is based on the number of students. 125. In terms of the institutional set up for the implementation of capital transfers, there has been a significant trend to remove the implementation of capital grants and capital budgeting from ministries of planning or economy and to integrate them with the rest of the budget process in ministries of finance. This has been an imperative result from the need to coordinate all aspects of budgeting. Despite that trend, 33 Matching arrangements are very commonly used, but they can raise some liquidity problems for low income subnational governments. This latter can be addressed by setting matching rates that are adjusted for the fiscal capacity of subnational governments. 34 Sometimes these approaches are combined. For example, an allocation of resources is provided by formula, but the awardees still have to through an application process and submit proposals in order to be able to use the funds. 56 countries often retain the vehicle of a Public Investment Program (PIP) but integrated into a Medium-Term Expenditure Framework (MTEF) or multi-year budget that covers the entire budget. Box 13. International Practices and Examples in the Design of Capital Transfers International experience with the design of capital transfers shows that a large variety of approaches are used. Most countries use some form of capital transfers in support of subnational governments for specific sectoral expenditure areas such as roads, water and sewerage treatment plants, transportation, housing, education, health, and so on. Country experiences vary along several dimensions. Regarding the mechanism used to allocate capital transfers, country experiences vary from ad hoc allocation decisions to formalized approaches using pre-established formulae. Similarly, in terms of the degree of flexibility in the use of capital grants, country experiences vary from the least flexible “project -based grants” to unconstrained funds provided as part of general revenue transfer. The variety of approaches observed make it particularly difficult to generalize and extract lessons useful for just any country trying to establish or reform a system of capital transfers. At the risk of oversimplifying, the typical country has a variety of capital transfers which are closed-funded in the national budget, which have earmarked funds within specific capital expenditure categories and require some level of matching funds from subnational governments, where the funds are allocated either by an objective formula or on a specific project basis based on qualifying proposals. In Australia the specific purpose transfers for capital purposes cover many areas including education, health, social security and welfare, housing, and roads. These transfers are administered usually by the relevant ministry of the federal government, and their design includes typically a formula used for the distribution of funds and an application process to pre-qualify. For example, capital transfers for education are administrated by Department of Education and those for roads by Department of Transport and Regional Services. In the case of education, capital grants are destined “government schools” and “non-government schools”. In the government sector, the Commonwealth’s capital funds are pooled with those of the state and the attribution of Commonwealth funding to individual projects is largely nominal. In the non-government sector, the allocation of Commonwealth fund is really based on projects assessment. The fund allocation among states for government schools is based on each State’s/Territory’s share of government school enrolments. In the non-government schools, the allocation is based on the number of non-government students enrolled in each state or territory. In the case of capital grants for road construction, one main program administrated by Department of Transport and Regional Service (the Black Spot Program) provides financial assistance to improve the physical condition or management of locations noted for a high incidence of crashes involving death or serious injury and it is run on the basis of individual project selection. The second program, which is formula-driven provides capital transfers for road construction to state and local governments. The formulas include a variety of factors such as population and road lengths. In Canada, capital transfers have been relatively small by comparison to the overall transfers (including the Health and Social Transfer, the Equalization transfer, and Territorial Formula Financing). Traditionally, some capital transfers are included in the Federal budget to share in the costs provincial and local infrastructure projects, such as highway construction by Transport Canada (the department of transport in the Federal Government of Canada). The Canadian Parliament also supports infrastructure financing through the Strategic Infrastructure Foundation, which matches private and public funding for large infrastructure projects. There is also funding available to Provinces and Territories to support affordable housing for low- and middle-income inhabitants. The Green Municipal Enabling Fund and the Green Municipal Investment fund also support energy and water efficiency projects. Funding is generally agreed to on a multi-year basis (most often for five years) and the bilateral agreements between Federal, Provincial, and Territorial governments. 57 In Mexico, the main capital transfer is the Contributions to Social Infrastructure Fund (FAIS), which is divided into two sub funds between the state and local. Capital transfers are provided for projects in drinking water, sewerage, drainage, and latrines; electricity in rural areas; basic health infrastructure; basic educational infrastructure; housing, rural roads, and productive infrastructure. Resources are distributed according to the relative share of each state in the extent and depth of poverty at the national level, measured by the Global poverty rate, which is the weighted sum of five indicators including income per capita, educational level, living space, drainage, and availability of electricity for cooking. In Brazil, the most important capital transfer is the State Investment Fund (FINEST), which has the objective of reducing regional inequalities by encouraging states that have HDI below the national average to invest in physical infrastructure for transport, energy, communications and sanitation. The Fund for the Development and Maintenance of Basic Education (FUNDEB), is distributed among the states according to its share of students, and 40 percent of the funds can be used for the building and maintenance of schools, and equipment. Source: Martinez-Vazquez (2000) and Zeikate (2002). 126. Finally, as in the case of most other transfers, incentives matter in the design of capital grants. One of the overriding concerns is that the funds indeed be used by subnational governments not only for the purposes intended but also that these funds be used in addition to those that the subnational government would have spent in the absence of the capital transfer. Achieving this “additionality” or “maintenance-of- effort” in capital transfers is not an easy task. This may be helped by earmarking the funds to be spent on capital investment of a certain type and by requiring subnational governments to match a particular share of the total capital expenditures. These measures contribute to the use of funds as intended by the central authorities. In addition, care must be given to avoid rewarding those subnational governments that have made clearly voluntary decisions to spend less on capital infrastructure and more on other types of expenditures. Otherwise, perverse incentives will be set in motion so that all subnational jurisdictions will wait for central handouts. Capital transfers must avoid creating the expectation among subnational governments that they will be the only source, or even the main source for capital infrastructure needed in the provision of decentralized public services. Capital grants should not be a substitute for using surplus funds and prudent borrowing policies under the “golden rule” (that borrowed funds can only be used for capital investment purposes). 5.2.2.5. Performance-based grants 127. Performance-based grants or “performance-based grant systems” (PBGS) have been suggested by experts and increasingly used in some countries as an alternative to specific earmarked and block conditional grants. 128. In general, PBGS incentivize improvements in subnational government performance and service delivery by linking performance in pre-determined areas with access to funding and with the amount of funding. The basic idea is to move away from ex-ante controls intrinsic to most conditional grants (be specific or block grants) to a system based on providing performance incentives coupled with ex-post monitoring and assessments carried out based on certain “agreed performance measures.” Essentially, there are three ways in which incentives can be provided: 1) by granting access to the grant scheme based on performance; 2) by increasing/decreasing the grant amount based on performance; 3) by allowing more discretion in the use of the grant resources based on the performance improvements. 58 129. In many cases, the performance measures—always linked to actions for which subnational governments are genuinely responsible-- are related in institutional dimensions, such as in revenue collection, planning, budget execution, accountability, financial management, or governance. But performance also refers to service delivery as measured by a variety of outputs and outcomes. Typically, the access to the program is linked to meeting some minimum conditions. 130. PBGS have been used in OECD and also developing countries in a variety of areas. For example, Australia used it in the program “Compliance with National Competition Policy to promote economic growth” to promote regulatory reform at the state level. In Canada PBGS has been used to promote equal access to health facilities; in Italy, PBGS was used to incentivize improved financial performance of subnational governments by topping up the European Union structural funds; in the UK, PBGS have been used in “Local Public Service Agreements” with significant additional funds provi ded for good performance on previously agreed indicators; and in the US, PBGS was used in the federal program for the states’ “Race to the Top” to incentivize reforms in certification and evaluation procedures of teachers and principals. In developing countries, PBGS have been utilized to improve service delivery in education (Uganda), health facilities (Philippines), water service (Tanzania) and the workings of subnational institutions (planning processes, compliance with procurement regulations, timely accounting, quality of audit, etc.) in many countries including Kenya, Bangladesh, and Indonesia. 131. The most delicate design area in PBGS lies in the specification of the performance measures. First, it is important to note that the socio-economic eligibility criteria can be very important to ensure that richer subnational governments—and therefore a priori already better performing--do not capture all the available transfer funds. Performance measures should include only indicators that are under subnational government control and should be employed only for activities that are completely decentralized. Because of the confounding impact of outside factors, such as per capita income or education levels of households, the focus of most PBGS has been on institutional process (such as budgeting or reporting) and on intermediate output indicators, which can be accepted as being controlled by subnational government actions. These types of difficulties have limited their application to more public sector areas. 5.3. Subnational borrowing and fiscal rules 132. Another critical element of fiscal decentralization architecture is related to subnational fiscal discipline. Subnational borrowing and fiscal rules can potentially play an important role in safeguarding macroeconomic stability and sustainability of subnational public finances. 133. A number of factors create adverse incentives for subnational governments to overspend, undertax, and/or borrow excessively – all root causes of fiscal imbalances. These excesses can compound fiscal profligacy by the central government, or undermine efforts to ensure fiscal discipline, thereby jeopardizing macroeconomic stability and, possibly, debt sustainability35. • Common pool problem: In an integrated national economic and financial space, the costs of fiscal indiscipline by one or more subnational governments (e.g. an increase in interest rates) are likely to spill over to the others. Moreover, each subnational government will have little incentive to save a windfall in shared revenues, as it may fear that it could lead to future reductions in its share, in favor of higher-spending jurisdictions. 35 Adopted from Teresa Ter-Minassian, Fiscal Rules for Subnational Governments: Can They Promote Fiscal Discipline? in OECD Journal on Budgeting – VOLUME 6 – No. 3 – ISSN 1608-7143 – © OECD 2007 59 • Moral hazard risk: It is often created by a history of bailouts of subnational governments by the central government. In such cases, the incentives for fiscal responsibility are weakened, as the costs of indiscipline are transferred to the national budget. Moreover, even an explicit central government’s commitment to a no bailout policy may lack credibility, if local governments are responsible for the provision of essential public goods and services, and a disruption in such provision is likely to have strong political and social consequences. • Inadequate transfer system and large vertical gaps: If subnational governments differ substantially in their ability to meet centrally mandated standards in the provision of the public goods and services they are responsible for, and transfers from the center are inadequate to bridge these gaps, pressures to run deficits may mount, especially if the subnational governments have access to bank or other forms of credit (e.g. supplier credit, or indirect borrowing through enterprises they own). • Fiscal autonomy36: Many subnational governments enjoy (to varying degrees) constitutionally established autonomy. This may prevent the central government from setting and enforcing effective budget constraints (borrowing limits, or balanced budget rules) on them. The greater the legal or even the de facto autonomy of subnational governments, the lower the probability that top- down constraints may be feasible and effective. 5.3.1. Subnational borrowing 134. All decentralized systems need to address the issue of long-term financing for the capital infrastructure expenditure needs of subnational governments. This is so because typically a considerable share of capital investment responsibilities has to be assigned to the subnational level. On the other hand, most subnational governments, in developing and developed countries as well, are not able to finance their capital investment responsibilities out of current savings. The same is also true for public utilities, when (and if) they are decentralized. These companies typically also lack the necessary funds for rehabilitation, maintenance and expansion of their capital stock. 135. Beyond the use of capital transfers, discussed already above, the other practical solution to this financing problem is for subnational governments (and their public utilities) to borrow the necessary funds for new investments in capital infrastructure, or for the rehabilitation of existing infrastructure if that is the case. On the other hand, subnational borrowing often has been seen as an activity full of risks and for that reason highly undesirable. 136. Despite those concerns, the overall consensus among fiscal decentralization experts is that subnational government borrowing for justified and needed long-lived infrastructure is both efficient and equitable. Borrowing is efficient because it allows subnational governments to make large lump-sum payments in order to acquire the necessary infrastructure and capital equipment for the provision of public services. In short, borrowing solves the problem of liquidity or the fact that current savings are inadequate for financing discontinuous capital investment needs. Borrowing is also efficient and equitable because it allows local governments to match the timing of consumption of services with payment for those services. Having one generation of taxpayers pay for the capital equipment and posterior generations consume its services free will not be fair. 36Despite the autonomy in fiscal policy it is generally accepted that the federal government ought to have some monitoring and control or implement mechanisms of synchronization of fiscal and debt policies to face macroeconomic shocks and to ensure individual and aggregated fiscal sustainability 60 Figure 7. Rationale for subnational governments access to credit Source: Own elaboration Two general forms of credit-based financing are generally available: bond issuance and financial intermediation. 137. The first is the direct access to capital markets by subnational governments by issuing bonds. The second is borrowing from financial institutions. Bond issues as the main means of financing long-term capital investments is unlikely to be an important option for many regional governments and for most local governments in countries in transition and in developing countries because of the lack of capital markets. Furthermore, bond issuing is a harder alternative for smaller local governments. An alternative to bond issues that has been used effectively in Western European countries in the past is the creation of a financial intermediary or a financial intermediation program that allows all local governments, especially those with no direct access to capital markets, to borrow based on selective banking criteria. 138. An important advantage of this approach is that these financial intermediaries can reduce the cost of borrowing for smaller governments by spreading the risks among many of them. It is also easier for a larger borrower to acquire a good reputation as a debtor and therefore be able to borrow funds in the capital market at lower costs. These lower costs can be passed on to local governments. Other advantages of this approach include the ability in some cases to combine technical development assistance with lending activities and also to facilitate central government intervention through the supply of funds. 139. However, the creation of a financial intermediation program also carries risks, such as political biases and abuses, that need to be assessed carefully. The experience with this type of financial intermediary in Latin American and other developing countries has often turned disappointing because of direct involvement by the central government’s mixing soft loans and political objectives with the strict lending banking criteria of financial intermediaries. There are lessons to be learned from the experience of other countries both for errors to be avoided and successful features to be imitated (see Box 14). 140. However, it should be clear that there is no intrinsic superiority to any particular approach to financing subnational government long-term capital investment needs. Direct borrowing from private commercial banks and other financial institutions, international lending programs, bond issues, and the creation of local government development funds are all desirable alternatives that need to be considered. 61 141. For the bond issue option, it is generally desirable that the Ministry of Finance or other designated agency act as a register to ensure that the bond issues conform with standard specifications to provide safety to investors and that subnational governments comply with borrowing limitations imposed by the central government. To facilitate bond issuance by subnational governments, it may be desirable to support the establishment of an insurance scheme for such bond issues. However, in no case should the central government act as guarantor of the bond issues; otherwise serious moral hazard issues would unfollow. 142. Public utilities or local governments responsible for public utilities should be allowed to issue a variety of bonds known as “revenue bonds.” These are non-guaranteed or limited liability debt to be paid from revenue proceeds associated with the public project the bond issue is supposed to finance.37 The funds needed to repay the bond are mobilized from setting public utility tariffs at full cost recovery levels. To guarantee payment these funds can be put in special escrow as they accrue. Box 14. Municipal Development Funds and Intermediaries* A “Municipal Development Fund” (MDF) is defined as a pool of money operated at a level above the individual local government primarily for investment in infrastructure. The MDFs are managed by different institutions such as banks or government agencies. These are the “municipal development intermediaries” (MDIs). With very few exceptions, Western European countries as well as Japan have had for decades a well- established MDI channeling investment credit to local governments. In many European countries these institutions were established also to provide a reliable outlet for private savings. Over the past several decades MDFs have spread rapidly through Asia, Latin America, and Africa. In many of these countries the financing of capital infrastructure has been combined with objectives for longer term institutional development. Objectives: The main objective of development funds is to mobilize resources from private lenders, central government, donor agencies, and local governments themselves, and make them available for investment in urban infrastructure. A second objective is to strengthen the operation capacity and efficiency of local governments by assisting them in the design, appraisal, and execution of investment programs, rationalization of programs moving away from ad hoc investment practices, and injecting rational criteria in the geographical and sectoral distribution of funds. Development funds are best suited to address the needs of smaller cities which tend to lack skilled administrators and lack access to capital markets. Development funds also have the attraction of offering a mechanism for “wholesaling” extensive programs for capital investment in infrastructure as opposed to having to micro-manage a myriad of small-scale projects. Management: The most common approach in Western Europe has been to introduce an autonomous institution with a legal and financial identity separate from the central government. However, there is great variance in the real degree of autonomy and the precise nature of the institution. Examples include: • Municipal development banks which are primarily concerned with financing municipal investment. These include Belgium and Danish Municipal Credit Associations, Bank for The Netherlands Municipalities, and the Municipal Bank of Norway. In all these cases municipal government representatives control the management board. 37 Revenue bonds are different from ordinary subnational government bond issues because the latter are considered “full faith and credit” debt, meaning that they carry an unlimited claim on the taxes and other revenues of the subnational government issuin g the bonds. 62 • Municipal or local government windows within institutions established to manage state- controlled pensions and insurance funds. This is the case in France, Italy, and Spain. In these cases, management is appointed by the central government, but municipal governments are represented in the decisions. The British Public Works Loan Board is 100 percent controlled by the central government, but half of the board draws membership from local governments. • Municipal “windows” of banks with a wider scope including mortgages and public works. This is the case of BANOBRAS in Mexico. Direct administration by the central government, the mode adopted by many developing countries. The central agency is the Ministry of Local Development or the Interior or, in some cases, the Ministry of Finance. Problems with some of these new funds included lack of capacity for sustained assistance, under-capitalization, poor loan repayment discipline. Funding: In Western Europe and Japan most of the activities of MDFs are funded by direct access to financial markets. However, the initial subscription of shared capital by either central or local governments, though not representing any significant resources for lending, has been important in establishing credibility and control. Additional resources are tapped by competing for private savings deposits. Other financial institutions (banks, insurance companies and pension funds) have been major sources of funds either through the purchase of bond issues or through directly negotiated deposits. In contrast, developing country MDIs have been largely financed with public funds. Even though most of these institutions have the power to issue bonds, most of them do not. Loans and conditions: MDFs lend money to local governments for long-term investment at preferential rates which cover interest and administration costs. In some cases, there are elements of grant or subsidized interest. These take different forms, such as matching grants attached to loans that vary with the repayment capacity of the local government or with the type of the project. Most often eligibility is unrestricted in which case the allocation of funds depends on the bids of individual authorities. In some cases, local governments, especially large units, are given a maximum quota. Security and Debt Service: Assessing debt service capacity of local governments is one of the most difficult aspects of managing a MDF. In most cases the policy is to rely on the statutory limitations established in the law, stating a maximum ratio of debt or debt service for local government revenues. However, in the case of self-liquidating investment, as for public utilities, the limitation depends on the internal financing viability of the enterprise and not the local government per se. There are exceptions to this rule. For example, the French Caisse de Depots bases its lending to local governments on financial forecasts, not on debt service ratios. Source: Based on “Municipal Development Funds and Intermediaries,” by Kenneth Davey, Background Paper for the 1988 World Development Report, The World Bank, July 1988.) 5.3.2. Fiscal rules38 143. Fiscal rules are a set of institutional constraints on policymakers’ decision-making discretion. Such rules may be imposed on sub-central governments by a higher level of government, or sub-central governments may adopt them themselves where constitutional arrangements grant them the autonomy to do so. One of the key motivations for subnational fiscal rules is their positive impact on economic stability and long-term fiscal sustainability: Fiscal rules can help prevent subnational fiscal autonomy from undermining sustainability at the macro level for example through excessive subnational borrowing. Where 38 Parts of this section are based on Martinez-Vazquez and Vulovic (2017) and Martinez-Vazquez and Civelek (2019). 63 they are allowed access to capital markets, sub-central governments without significant revenue raising autonomy are more likely to be perceived by lenders as borrowers that are implicitly guaranteed by central government. The costs to the rest of the country of profligate fiscal policy can thus arise through the moral hazard of bail-outs from central government, rising overall borrowing costs and higher and more volatile public spending and tax rates. The are also other motivations behind subnational fiscal rules that are related to increasing aggregate and allocative efficiency. 144. There are two fundamental reasons why subnational governments may be undisciplined regarding borrowing. First, the design of the fiscal decentralization system may be—typically is—faulty. While most fiscal decentralization reforms have devolved significant expenditure authority, they have done so much less in terms of revenue authority. This form of asymmetric decentralization leads to large vertical imbalances, which are routinely closed with high degrees of local government dependence from central transfers. This in turn, through the common pool problem, leads to excessive local spending and lower tax effort since transfers represent little political cost to subnational authorities, and also leads to moral hazard and expectations of bailouts for excessive borrowing. Second, even when no such defects are present in fiscal decentralization design, there are simple political economy issues at work that must be recognized. In simple terms, borrowing allows even huge increases in local spending, the cutting of ribbons for infrastructure projects by subnational officials, etc., while it postpones most of the payments to future times. Subnational officials seeking votes may see additional borrowing and spending as an expedited way to gain political support, while the (political) costs are shifted to someone else in the future. 145. Thus, ensuring disciplined and responsible subnational borrowing behavior generally requires more than good decentralization design. It also requires explicit rules, monitoring and oversight regarding subnational debt. 146. The question is how to provide a legal framework that allows subnational governments flexibility to pursue optimal service delivery at the same time preventing undisciplined or irresponsible behavior. The main types of borrowing and fiscal rules used in the international practice are described below. But beyond those fiscal rules, some countries have been used other supporting institution for monitoring spending and borrowing behavior and helping maintain fiscal discipline. Two of these institutions are “fiscal councils” (Box 15) and “rainy day funds” (Box 16). Box 15. Fiscal Councils The most recent innovation in fiscal governance over the last decade has been the introduction of fiscal councils. In some cases—like in the cases of Hungary and Nigeria-- fiscal councils have been established as a part of fiscal rules package. In all cases, developed and developing countries, fiscal councils have been adopted as a more or less independent authority to monitor and help enforce fiscal rules. The original idea had been whether a truly independent authority --similar to the role played by central banks regarding monetary policy-- should be put in to control government debt and deficits; the main argument being that most governments are time inconsistent regarding their short-run fiscal policy objectives and long-term fiscal stability performance. Many existing fiscal councils were created with a main focus on central government performance but more recently the scope of has been extended to also the subnational level. What is the effectiveness level of fiscal councils? This institutional innovation is fairly new, but the preponderant evidence is that they can play a significant role in improving fiscal discipline and overall performance. This is especially the case when they are provided with more independence and authority. The evidence regarding the impact on subnational borrowing and fiscal discipline is still limited. 64 However, clearly, the combination of public reports and high media impact of fiscal councils has good potential to lead to better fiscal outcomes at the sub-national government level. But note that even though most existing councils make some forms ex-ante and ex-post policy assessments and fiscal sustainability analysis, practically none of them go beyond the evaluation of current policies and forecasts; in particular, they lack any legal authority to determine national debt or deficit levels, which are still reserved –unlike the case of independent central banks--- to elected authorities. Source: Martinez-Vazquez and Civelek (2019) Box 16. Rainy Day Funds The “rainy day fund” is a subnational fiscal institution which has been around for a long time but which has attained special attention in the aftermath of the Great Recession. Practically all states in the US have adopted “rainy day funds” to protect their budgets and spending against unexpected economic shocks and uncertainties and to minimize the disruptions caused by economic fluctuations and sharp declines in revenues. Their effectiveness in the past has depended on the size of the shock and the types of expenditures protected. There is no consensus on the optimal amount of rainy day funds, but the recommendations following the 2008 crisis increased up to 15 percent of annual revenues, against the suggested rates of 3 percent to 5 percent in the 1980s. Source: Martinez-Vazquez and Civelek (2019) 147. Following Ter-Minassian and Craig (1997), most existing borrowing and fiscal rules fall into four general categories:39 • Administrative Approach. This approach gives power to central authorities to directly control subnational borrowing and debt through different instruments, including debt ceilings, prohibition of external borrowing, and prior approval of conditions for any new debt. The direct control of central government on subnational government borrowing is more frequently seen in unitary countries –and in those at early stages of decentralization. It provides tight macroeconomic stability tools, but it weakens fiscal decentralization. For example, Mexico combines hierarchical and market controls. That is, in addition to administrative controls, Mexico uses financial sector regulations—essentially a system of credit ratings that add another layer of prudence to all subnational borrowing activities. Chile can also be considered as part of this group, which in the past has prohibited subnational borrowing. • Cooperative Approach. Neither the law or higher-level governments determine subnational borrowing and debt regulations. Instead, those regulations are determined through negotiations and dialog between the central and subnational governments involving deficit targets and debt ceilings. Dialog raises general awareness at all levels, but negotiations can break down with negative consequences for budgetary discipline. South Africa is a prominent example of this approach. • Rule-based Approach. Rather than resting on central discretion or voluntary agreements, this approach relies on compliance with clearly stated ex-ante rules that subnational governments need to follow when they borrow. To restrict subnational borrowing, fiscal rules are specified in a legal 39Actually, many countries use features of different regimes. Countries also frequently change subnational borrowing regimes; Chile has been among the exceptions of maintaining the same regime, maintaining the prohibition of subnational borrowing. 65 framework listing in different combinations: debt ceilings or limits on total borrowing, ceilings on debt service expenditure (for principal and interest), the “golden rule” (funds can only be used to finance investment on infrastructure), rules on who can be a lender, whether or not lenders can be foreign entities, and so on. A clear goal is to relate the level of borrowing to the repayment capacity of the borrower entity. A second goal is to make subnational fiscal outcomes predictable. Because most often the rules are specified in the national legal framework, they are known as centrally- imposed rules. However, it is possible for subnational governments to also have self-imposed rules. While it may seem relatively straight forward to monitor legal limits of total debt, or limits on debt service expenditures, these rules are not always effective because subnational government may use different procedures to go around them (Box 17). The rule-based approach has the clear advantages of transparency and be free of unnecessary bureaucracy or long negotiations; on the other hand, it can be inflexible in financial crises. In practice, this is the most commonly used approach having been adopted in countries like Brazil, Mexico, Argentina, Bulgaria, Colombia, Czech Republic, Hungary, Poland, Romania, and many others. • Market-based Approach. Under this approach, the central government does not regulate subnational government debt limits but rather relies on financial markets to regulate and restrict subnational borrowing. Subnational governments are free to decide on the source of the debt, its terms and conditions, and the amount that they want to borrow, as long as the financial markets institutions allow it. This system presumes the availability of accurate information and indicators on the subnational governments’ ability to repay the debt, while there should be a clear understanding of no bailouts by the central government. All these requirements mean that this approach is used just a few highly developed countries, prominently in the USA. • Mixed systems. Even though we have emphasized above the main differences between approaches, in reality many countries, as already pointed out, combine aspects of the different approaches to build up their own systems. Box 17. Circumventing borrowing rules Subnational governments can be resourceful and creative in circumventing legislated borrowing rules. Some of these common practices include: • Subnational governments may circumvent rule-based borrowing regulations by labeling current expenditure as capital expenditure. • They can use subnational government financial vehicles, which have extra-budgetary status, such as government-owned enterprises. Subnational government may borrow using these entities that are mainly out of debt ceilings determined by fiscal rules. • Similarly, using local government financial vehicles the subnational government may borrow to fund some projects that should be funded by the subnational government own budget. • The rule-based approach may lead to accumulation of short-term debt that is often not clear whether it is included in debt ceilings or not. 66 • Subnational governments can accumulate considerable payment arrears to providers, central government agencies and even employees; these may not be readily apparent when cash- based accounting is used --as opposed to accrual) Source: Based on Ter-Minassian (1997a) Are there better performing or preferable borrowing and fiscal rules? 148. One last question is whether there are better performing rules. In popularity or common use internationally, the rules-based approach would appear to be the winner. Many more countries have adopted this approach. Its main attraction is that it combines transparency with strong indirect central government control. 149. Generally speaking, fiscal rules can be divided in two main categories: procedural and numerical rules. Procedural rules aim to enhance transparency, accountability, and fiscal management. They typically require the government to commit up front to a monitorable fiscal policy strategy, usually for a multi-year period, and to routinely report and publish fiscal outcomes and strategy changes. Numerical fiscal rules refer to specific quantitative targets. They are intended to impose permanent constraints on fiscal policy, typically defined in terms of an indicator of overall fiscal performance (such as the budget balance and/or the public debt). 150. However, each approach encompasses multiple dimensions, and clearly none of the approaches a priory dominates all those dimensions. In addition, the statistical analyses that have been performed so far in the economics literature show that no particular borrowing regime dominates the rest in terms of effectiveness and overall fiscal discipline results. Nevertheless, the more general rules, including the “golden rule” and limits on total debt and debt service appear to be quite effective. On the whole, effectively successful fiscal discipline at the subnational level requires going beyond subnational borrowing and fiscal rules, focusing on preventive measures (such as the streetlight system introduced in Colombia and Mexico), institutionalizing monitoring (via fiscal councils, for example), introducing enforcement measures (for example, by using the intercept of other subnational funds such as transfer to ensure compliance with repayment terms) , and, if needed, also the application of sanctions (including the possible application of bankruptcy proceedings and administrative interventions of subnational governments). The strength with which the rules are applied, comprising the monitoring and enforcement mechanisms, ultimately could be more important than the rules structure. Table 2. Types of subnational fiscal rules Category of Type of rule Description Examples of countries rule Numerical Debt rule Limit on total outstanding Argentina, Brazil, Italy, Japan, debt or debt service with Spain, Lithuania, Romania, Poland, relation to local revenues. Colombia, Peru, Colombia, Korea, Canada, Czech Republic, the Netherlands, Portugal, Turkey Budget balance Specific targets are set for Current budget balance: rule subnational balance. Germany (local), Japan (local), Sometimes the rule takes a Netherlands (local), Italy (state), form of “golden rule” when France (local), New Zealand (local), borrowing is only permitted Sweden (local), Switzerland (local), 67 for capital expenditure Canada (local), Finland (local), (current balance rule). Norway (local) Overall budget balance: Canada (state), Czech Republic (local), Denmark (local), France (local), Korea (local), Portugal (local), Turkey (local), Spain (local and state); Poland (local); Peru (local) Expenditure rule Expenditures may not grow Germany, Portugal, Korea, Turkey faster than some determined (all local) rate Revenue rule Floors or ceilings are set on Not used frequently at subnational government revenue level; sometimes in relation with natural resource revenue in the context of stabilization funds. Procedural Legal approval Local councils are required Canada, Switzerland, USA to approve borrowing for individual projects Source: Authors’ elaboration based on Martinez-Vazquez and Civelek (2019) and Sutherland et al. (2015) 68 6. Fiscal accountability and transparency 151. Accountability is not an automatic outcome of increased discretion of the local governments. Specific mechanisms should be designed to ensure that citizens and higher officials are able and willing to hold local governments accountable for their discretion. The notion of fiscal accountability relates to issues related to public accountability, through well-established relations between the central government and subnational governments, properly designed budgeting process, audit, control and evaluations, as well as social accountability through transparency and citizens’ participation. 6.1. Evolving role of central government in decentralized countries 152. The impact of decentralization reforms on the central government is often underestimated. By many central governments, the decentralization process may be perceived as a threat. As reported by the OECD (OECD, 2019), subnational governments are often facing strong resistance from the central government agencies. This is due to several reasons: (i) national government tends to have limited trust in subnational government competency and accountability for failure; (ii) decentralization may represent a loss in terms of public employees and control on public affairs; (iii) decentralization requires profound changes in organization, roles, practices, culture and skills within the central government itself. 153. The new roles of the central government, some of which are related to public accountability, include for example: • Departure from the implementation towards a more strategic role. In the decentralized environment, the focus of the central government is increasingly on putting in place the strategies, framework conditions and regulations, incentives – rather than in the implementation of policies. It involves a shift from a direct role in service delivery to one of enabling, advising and assisting, ensuring consistency, facilitating the work of local governments and sometimes helping share good practices across local governments. This requires building new capacity at the central government level, able to cope with these new functions, which cover a large area of sectors; • Supporting role for subnational governments especially in the capacity constraint environment. The central government has important role to play in ensuring that decentralization does not widen disparities in terms of access and quality to the public services provided by subnational governments in their new functions. This can be achieved with appropriate instruments aimed at monitoring, diagnosing and improving the performance of local services. Another challenge for the central government is linked to human resources management, to make sure that decentralization does not generate a two-speed civil service. Central governments have a crucial regulatory and co- ordination role to play in ensuring that subnational governments have the necessary autonomy and flexibility in this area. Finally, central governments can establish adequate co-ordination mechanisms across different levels of governments and promote cooperation and dialogue. • Strengthening of the supervisory role of the central government. The central government also needs to renew its supervisory and monitoring role vis-à-vis subnational governments, especially in legal and fiscal matters. Control should be based on posteriori rather than a priori control. Financial audits are necessary to assess the quality of financial reporting and the reliability and accuracy of financial information and management provided by subnational governments. 6.2. Budget process for regional governments 154. The basic principle of national budgeting is that each level of the government should have its own budget to cover its own sphere of activity and responsibility. The budget for subnational governments is heavily influenced by country-specific circumstances, in particular: • Scope of Autonomy: Local jurisdictions without expenditure and revenue authority and without discretion for use of funds are almost certainly hampered in their capacity to effectively budget. The scope of centralization and decentralization is, therefore, an important antecedent factor in local budget processes. Decentralization implies meaningful choice and the ability to tailor local public sector policy and service delivery to the needs of local populations, as well as incentives for an effective, efficient, and locally accountable public sector. At the same time, mechanisms must be established to effectively link national and subnational planning and budgeting processes. This coordination becomes a defining element of subnational budgeting. • Coordination of Multitiered Institutions and Actors: Effective local budgeting requires central and regional governments to promote the development of local jurisdictions as effective governance institutions. Higher-level jurisdictions are critical in establishing procedures and standards for improved local operations and transparency and for promoting the accountability of local officials to local constituencies (Smoke 2000). Central governments must conduct analytic studies to monitor and evaluate the decentralized system, determine tax effort, track subnational budgets, evaluate fiscal alternatives, establish and administer transfer systems, regulate local borrowing, establish fiscal information systems, and establish audit procedures. In addition, they must establish personnel and environmental regulations, qualifications for public employees, minimum service standards, mechanisms to resolve intergovernmental disputes, and sanctions (World Bank 2000). • Tools: Subnational governments, as much as central governments, require a well-structured PFM system to help ensure appropriate and accountable use of public resources in pursuit of developmental goals. Shah (2007) provides a detailed treatment of the elements of subnational public financial management, including basic accounting and reporting, cash management, procurement, debt management, internal control and audit, as well as external audit and performance evaluation. 70 Box 18. Link between National and Subnational Budgets Effective coordination among levels of government is critical. The link between national and subnational budgets is strongest in centralized systems in which subnational governments function as administrative divisions of the central government. Central ministries establish priorities, and often subnational budgets are centrally approved. In decentralized systems, national and subnational budgets may not be formally coordinated (as in the United States) or these budgets may be integrated into one framework (as in Germany). In the United States, the national government uses fiscal incentives to entice states and localities to conform to national policy, and the national constitution and statutory legislation somewhat limits subnational units’ budgetary purview. Mechanisms of direct budgetary coordination are absent. Germany’s 1969 Law on Budgetary Principles created a framework for coordinating budgetary processes and establishing uniform budgetary principles across levels of government. General provisions establish requirements for fiscal year, gross estimates, comprehensiveness, spending authorization, efficiency, and cost-effectiveness. Specific provisions establish budget preparation means, the accounting structure, budget classification, auditing, execution and reporting, and the role of special funds. Source: Shah, Anwar. 2007. Local Budgeting. Washington, DC: World Bank. © World Bank. https://openknowledge.worldbank.org/handle/10986/6669 License: CC BY 3.0 IGO. 6.3. Audit, control and evaluations 6.3.1. Audit and control40 155. An internal control and audit frameworks aim at improving financial and administrative management capacity by limiting fiscal behaviors that result in waste, misallocation, and corruption. In order to achieve efficiency and effectiveness at the subnational governmental level, internal control and audit should be among the key components of a fiscal decentralization program. 156. Internal control and audit are key components of public financial management systems for increasing efficiency and effectiveness in local government operations. An effective, efficient, transparent, and rules-based PFM system is an essential tool for a government in the implementation of a fiscal decentralization program. Within the PFM framework, internal control and audit functions support fiscal decentralization process by promoting transparency and accountability in the use of public resources, ensuring allocation of public resources in accordance with citizens’ priorities, and supporting aggregate fiscal discipline. A well-functioning internal control and audit system provides all stakeholders with broader outputs and outcomes as it seeks to deliver better guidance to the management by mitigating and eliminating risks, developing effective strategic development plans, removing inherent errors and inconsistencies within the system, and strengthening integrity through multiple-checks. Thus, effective decentralization strategies require that the subnational government strengthen its internal control and audit instruments while devolving more power to local governments. 157. In addressing control and audit reforms, subnational government practitioners need to account for the different dynamics of internal control and audit. Both entail careful consideration and demand a series 40For more information please see: Baltaci, Mustafa; Yilmaz, Serdar. 2006. Keeping an eye on subnational governments: internal control and audit at local levels (English). WBI working paper series. Washington, DC: World Bank. http://documents.worldbank.org/curated/en/458411468178144890/Keeping-an-eye-on-subnational-governments-internal-control- and-audit-at-local-levels 71 of subsequent actions. It is important to note that internal control and audit systems are not a complete or magical solution on their own. They do not necessarily resolve all challenges in a public financial management apparatus. Instead, to improve governance structures and accountability frameworks necessitates a broader strategy encompassing the whole spectrum of PFM issues, which include planning, budgeting, accounting, procurement, public debt, asset management, reporting, external audit and legislative oversight. 6.3.2. Evaluation function41 158. After government programs have been in place for a while, there is always a need to ask about their effectiveness. Have the goals of the program been achieved? Have the objectives of the grant been obtained? Is there a need to change part of the design or even discontinue the program? Has the chosen allocation approach resulted in the expected distribution of resources across state governments? In some circumstances, conditional grants and other policy programs achieve all they were designed to achieve. In other circumstances, they achieve very little (if anything). But, how does one know? There is wide agreement among experts that all transfer programs should be subjected to rigorous evaluations. 159. And an increasing number of countries have put that into practice. For example, in Australia, the evaluations of transfer programs, but also of all other federal government spending programs, have been carried out periodically since the budget management reforms of the 1980s. In the United States, routine evaluations of federal programs are carried by the GAO (the US Government Accountability Office). The most important transfer program in the European Union, the Structural Fund is, with other similar programs, evaluated routinely since the mid-1990s. Currently, no new transfer programs can be implemented in the European Union without a previous evaluation. Some countries, like Sweden, go back many decades with routine evaluations of transfers and other central government spending programs. 160. It is important to remark that evaluation is quite different from monitoring. In the case of monitoring the concern is with compliance with overall conditionalities, schedules, uses of inputs, and meeting financial requirements. With monitoring, ultimately, the central government must be satisfied that the conditions of the grant are being met before further payments are made. The most efficient place for the final certification to be made is in the agency from which the future payments have to be authorized. In contrast, the nature of evaluation is to focus on comparing the intended objectives and the actual impact on the targeted clientele and whether the desired changes can be observed. 161. Typically, performing a good evaluation is not an easy task since it is complex to isolate the impact of the transfers or other policy program in general from other confounding factors that may also have a bearing on the observed outputs and outcomes. Essentially, the questions to ask are: what would have been the outcome or outputs in the absence of the transfer? Are the observed outcomes cost effective given the size of the transfers, or are there other more cost effective ways to achieve those outcomes? How are the incentives of the state governments affected? Have the transfers simply become a substitute for other funds that the states would have spent anyway, or do they instead provide additional financing and enhanced outputs and outcomes? Can the transfer program be redesigned to improve targeting of the intended groups? Is the evaluator independent enough and the methodology used credible enough? Are policymakers expected to act on the results of the evaluation, including the discontinuation of the transfer program? 162. What can be learned from the international practice in institutionalized form of evaluation? The institutional setting varies significantly, but leading countries in this field tend to share underlying principles and consistently focus on ex post evaluation of selected programs and projects. For example, Canada is an 41 This section is partially based on Martinez-Vazquez (2019). 72 example of a decentralized system where the center broadly regulates activity, while Colombia has a dedicated unit in the Ministry of Planning and Mexico has an independent agency mainly focused on the evaluation of social programs. And even though institutional arrangements differ across countries, some fundamental operational principles are shared in practice, including a general policy that regulates the evaluation activity and sets standards, the preservation of transparency and independence in the evaluation process, and the existence of a formal evaluation agenda. Box 19 presents several different experiences in institutionalizing the evaluation function. 73 Box 19. Experiences in institutionalizing an evaluation function for public policies In Mexico in 2000, the Mexican Congress passed a law requiring an annual evaluation of all the programs conducted by the Federal Executive Branch. The declared objective was to achieve greater transparency and accountability and to prevent political manipulation of public programs, In 2001, the Under Secretariat for Planning and Social Evaluation in the Secretariat of Social Development (Secretaría de Desarrollo Social, SEDESOL)was created, and in 2004, the Mexican Congress approved a Social Development Law that institutionalized the evaluation process and created the National Evaluation Council (Consejo Nacional de Evaluación de la Política de Desarrollo Social, CONEVAL). CONEVAL has since been coordinating the evaluation agenda of social programs, developing standards and terms of reference for different types of evaluations, generating or contracting external evaluations, and coordinating results-based monitoring and evaluation system. In Mexico, evaluations are also carried out in the education sector by the National Institute for the Evaluation of Education (Instituto Nacional para la Evaluación de la Educación, INEE). Colombia has achieved considerable success in the creation and strengthening of its whole-of- government monitoring and evaluation system (Sistema Nacional de Evaluación de Gestión y Resultados SINERGIA. It is one of the strongest monitoring and evaluation systems in Latin America, in terms of its overall credibility and utilization. SINERGIA also includes an ambitious agenda of evaluations, particularly impact evaluations. These ex post evaluations are generated to inform policy decisions and measure result achievements. Institutionally, they are centrally driven and coordinated by the Directorate for Public Policy Evaluation (Dirección de Seguimiento y Evaluación de Políticas Públicas, DPPE) in the National Planning Department (Departamento Nacional de Planeación, DNP) in coordination with the Ministry of Finance and the sector agencies. The system was constitutionally mandated in 1991 and formally implemented from 1994. The system always has had strong political support from the Office of the President, which is one of its major users. Operationally, it runs on an online sub-system for monitoring and reporting and has an annual evaluation agenda of around 15 impact evaluations Over the years, the system has increasingly been used by the government, as well as by civil society and the media. It is supported by strong staff technical capacities hosted in the DPPE. In Canada, the evaluation activity is more decentralized at the departmental and program levels, but under the regulations and guidance from the center. Specifically, the Centre of Excellence for Evaluation (CEE) under the Treasury Board Secretariat (TBS) provides functional leadership, including advice and guidance in the conduct, use, and advancement of evaluation practices across the federal government. Evaluation is understood as the systematic collection and analysis of evidence on the outcomes of programs to make judgments about their relevance and performance and to examine alternative ways to deliver them or to achieve the same results. It is meant to support accountability through public reporting on results, expenditure management, management for results, and policy and program improvement. The CEE fulfills the TBS's responsibilities under the 2009 Policy on Evaluation to provide functional leadership for evaluation across the federal government through advice and guidance in the conduct, use, and advancement of evaluation practices. The CEE carries out its responsibilities primarily by engaging directly with departmental evaluation units and evaluators and providing support on the use of evaluation findings in central agencies and across the government. As a functional leader for evaluation, the CEE engages with a variety of internal and external partners and stakeholders. The CEE mainly develops guidance materials and tools to support policy implementation in departments and agencies and advances evaluation practices, monitors evaluation capacity and policy implementation in departments, and fosters community and capacity development to support improvements in the capacity and competence of evaluation units across the Government of Canada. Source: World Bank (2018). 74 6.4. Transparency 163. Given the increased responsibility of subnational governments to deal with resource allocation and service provision, the importance of transparency at the subnational level has increased dramatically. Transparency at the subnational level is required in order to expose service delivery issues, which are a major concern in many countries. Moreover, subnational budget documents deal with items and expenditures that are particularly visible and tangible for citizens. In this sense, subnational transparency has the potential to provide answers to basic questions such as what happened to central government funds at the local level and whether plans to build a new neighborhood school, or carry out local road works were actually implemented. 164. Although there is a limited literature on subnational budget transparency, the following principles were developed by the International Budget Partnership and can help guise the design of the transparency system42: 1. Subnational governments should provide fully transparent and comprehensive information to the public regarding both budgeted and actual revenues and expenditures, through all four stages of the budget process (formulation, approval, implementation, and evaluation). 2. Subnational governments, like national governments, should be audited internally and externally, and external audit findings should be made available to the public within a reasonable time frame. A report detailing actions taken should be published within a set timeframe (laid down in regulations or law) by the auditor describing responses from officials and sanctions applied where necessary. 3. Where subnational governments have discretion over the use of funds, which have been transferred to them from other levels of government, or funds collected through their own revenue sources, opportunities should be provided for local assemblies (where these exist) and citizens to provide input into the budget process at multiple points from allocation to implementation and evaluation. 4. The executive’s budget at central and subnational level should contain detailed information (quantitative and narrative) related to intergovernmental transfers, including at least their size, all rules for receiving and spending them, their distribution, the rationale for this distribution, and, where applicable, the formula by which the distribution is calculated. Where states or provinces distribute funds to district levels, the same should apply. 5. The distribution of resources across subnational units should take account of equity considerations and should be driven by the need/demand for services, rather than exclusively “supply -side considerations” (i.e., the existing supply of infrastructure and personnel). The budget should contain narrative and quantitative information explaining how the need/demand for services is measured and its relationship to the distribution of resources. 6. Where funds are decentralized, the executive budget proposal or other publicly available and easily accessible central government documents shall contain narrative information detailing the responsibilities of each level of government for the delivery of the decentralized services, the public offices or public officials within each level of government that are responsible for specific services, and the fund flows associated with those responsibilities. 42 https://www.internationalbudget.org/wp-content/uploads/Subnational-Synthesis-Paper_final.pdf 75 6.5. Citizen participation 165. Social accountability mechanism for decentralized governments can be strengthened through platforms allowing for citizen participation. There are different models how this can be achieved. Some of the common examples include citizen monitoring committees and participatory budgeting. • Citizen monitoring committees to perform independent budget analysis and public expenditure tracking programs: Effective accountability can be instituted when different groups of citizens come together to monitor various aspects of the budget and expenditure process. • Participatory budgeting: Allows the citizens to present their demands in terms of allocation of budget through discussions and negotiations with the local government officials. This practice does not only ensure that citizens’ preferences are included in decision making but also allows the citizens to become active participants in community problem-solving. 6.6. Decentralization as means to address regional grievances 166. Historically, across countries, decentralization has been used as a means for addressing active regional grievances, typically using asymmetric fuller concessions for fiscal autonomy to the grieving regions. Such has been the case, for example, in Indonesia and Spain, as documented in the second accompanying volume with country case studies. However, the specialized literature on this issue does not provide strong support on whether decentralization works as glue or solvent of countries’ unity and fragility. Further fiscal decentralization may be an alternative to secessionism by providing a political protagonist role to some regions, but at the same asymmetrical decentralization arrangements favoring secessionist regions can also boost perceptions of discrimination and ultimately lead to more conflict in the longer run. The relevant literature on these issues is explored in Martinez-Vazquez et al. (2017). 76 7. Implementing decentralization reform43 167. Undertaking comprehensive fiscal decentralization reform faces lawmakers with the choice of either sequencing those reforms over time, or clustering the reforms in a short period of time by simultaneously adopting all, or the most important, pieces at once. A sequenced gradual implementation of reform is generally thought to be more desirable than a “big bang” approach because it carries substantially lower risks in terms of the information available at any time and the ability to spread the costs of the reform over time. A sequenced approach may also allow for some experimentation and being able to adjust to different initial capacities of subnational entities. However, gradualism and a paced implementation should never be interpreted as not requiring a comprehensive blue print for decentralization reform. Proper sequencing demands a roadmap with explicit plans and goals, accompanied by the institutional and administrative changes necessary to accomplish those plans and goals. 168. However, there may be political imperatives present that may justify a more hurried up approach to reform as in the case of significant transition in political regime. More generally, policy makers may lack flexibility to reform in a gradual fashion following an economic crisis. Going with a big bang approach to decentralization reform in these circumstances can be an acceptable risk, and which may generate acceptable outcomes, as witnessed by the relative recent cases of Indonesia and Poland (reviewed in the second accompanying volume). But a big bang, abrupt reform efforts may also lead to failure, as documented, for example, for some Lain American countries in Giugale et al. (2000). 169. The introduction of important reforms also requires adopting transition measures for their differential impact on subnational entities. Most fiscal decentralization reforms imply that there will be winners and losers. Politically or relatively economically powerful subnational governments may threaten the reform unless their concerns are addressed. Here, policy makers can select to go with a “hold harmless” approach, where no entity end up with less resources it had before the reform, or with a so-called “cold turkey” approach, where the new system is adopted no matter what, or, finally, with an intermed iate approach, where the reform are introduced over a period of time and losers are at least partially compensated. The political economy of reform in most countries often calls for some form of sequencing with compensation, if not a strict “hold harmless” approach. 170. One last dimension of any successful decentralization reform effort is the planning and implementation for monitoring and evaluation, which should lead to periodic fine-tuning of institutions and processes, as well as to the provision of technical assistance to subnational governments in those areas where they are underperforming. Carrying out successful monitoring and evaluation typically requires a unit, most likely located in the Ministry of Finance, with trained staff exclusively dedicated to this task, and an extensive information and data system that will enable proper monitoring and evaluation. 8. Lessons learnt on fiscal decentralization 171. There are several basic principles that should guide decentralization reforms in general to ensure accountability, transparency, and an efficient performance at all levels of government.44 In what follows those principles are listed and an attempt has been made to relate them to the Chilean case; however, 43 This section is based on Bahl and Martinez-Vazquez (2006). 44 See Bahl and Martinez-Vazquez (2006).. because the Chilean process is still fairly new and not quite defined, the comments will necessarily remain general: • Fiscal decentralization policy should be comprehensive. Ideally, the design of a decentralization program should address all of the components of the system even if not all components are implemented right away. Successful decentralization policies build on more than the traditional fiscal issues, including an electoral system and other forms of political accountability and a capable civil service. Piecemeal reform, encompassing only one element of the system, such as increased financing of local governments, is unlikely to deliver the benefits of decentralization. For Chile this means that the creation of the new regional governments in Chile will demand a detailed “white paper” and road map covering all dimensions of decentralization: fiscal, political and administrative decentralization, and within the fiscal decentralization dimension including the architecture of the assignment of functional expenditure responsibilities, the assignment of own revenue sources, the system of intergovernmental transfers, and last, the prudential access to borrowing for capital investment purposes. The temptation will be just to do part of the work at this time and postpone other aspects to a later day. However, that will likely imply uncoordinated, if not dysfunctional, decentralization, creating problems that will be hard to solve down the line, as the process becomes increasingly path-dependent. Nevertheless, the requirement that the design should be comprehensive does not necessarily mean that the full architecture needs to be implemented all at once. This latter is not the case. However, there should be a clear roadmap for when and how the different elements of the reform will be implemented, accompanied by a full political commitment to do so. • Finance should follow function, which means that the first step should be the assignment of expenditure responsibility to subnational governments, and only following that the assignment of revenue-raising powers and central government transfers should be determined. Known expenditure responsibilities facilitate the determination of the level of funding and the types of revenue instruments: what services should be financed by user charges, local taxes, conditional grants, and so on. On the implementation side, the decentralization of expenditure competencies needs to be carried out in the context of reformed national sectoral laws that support and do not hinder the decentralization reforms. Subnational governments must have the capacity to deliver the assigned services or else the expenditure assignments may need to be asymmetric and restricted to those that have these capacities. Notice that determining the required level of funding demands a determination of the cost of providing these services. This can be done by adopting a “retrospective” method of using some form of per capita (or per client) financial norms or physical service standards. For Chile this means that no decision should be made about the financing sources for the new regions until there is clarity on what functional responsibilities or elements of those responsibilities will be assigned to the regions and what the expenditure needs associated with those responsibilities will amount to. The exercise of assigning expenditure responsibilities to the new regions will be demanding because some of those responsibilities will be detached from the central government portfolio and some from the local governments’ portfolio. There will be also a need to review and update sectoral laws (on education, health, transport, etc.) in Chile as many of those laws and related regulations are likely to refer to the functional responsibilities of the new regional governments. • There should be clarity of roles between levels of government. Clarity, transparency, and stability are paramount for achieving the accountability at the administrative as well as the political levels 78 that efficient governance requires. Much too often, overlapping responsibilities accentuated by expenditure dumping or unfunded mandates combine with frequent legislative and regulatory changes to create a diffused and confusing system of expenditure responsibilities that compromise efficient service provision. For Chile, from this perspective, the new legislation on regions should strive to bringing clarity to the assignment of functional responsibilities for the very start in an uncompromising way. These assignments preferably will be permanent or at least not often changed. • A significant measure of autonomy for subnational governments on both the expenditure and revenue side is crucial for realizing the efficiency potential of decentralized government. On the expenditure side, this would require providing subnational governments with budget flexibility to decide—within limits—expenditure priorities, the service levels offered in accordance with local preferences, and the best means to deliver those services. On the revenue side, a minimum of subnational autonomy to set tax rates on at least one significant tax source assigned to the subnational level; allowing subnational governments the powers to raise revenue is a requirement for accountability. Revenue autonomy also facilitates addressing vertical fiscal imbalances and building up credit-worthiness at the local level. Generally, subnational governments should not be allowed the powers to define tax bases. The best vehicle for local autonomy is discretion in setting tax rates for some meaningful taxes. While autonomy should be explicit and well-defined, it must also be circumscribed in some important respects, first of all with respect to the access to borrowing by subnational governments. From a Chilean perspective, the fundamental question to be asked is why there is a need to further decentralize government in the country by creating new regional governments. The answers are likely to include the need to bring government closer to the people, increase the efficiency of public expenditures, to further democratize the country and to achieve more balance growth and distribution of resources across the different regions of the country. However, decentralization will not bring the country much closer to any of those objectives unless the decentralization process provides significant measures of autonomy and discretion to the new regional governments and not only on the expenditure side, but also on the revenue side of the budget. • Decentralization should go hand in hand with an intergovernmental transfer system that has elements of equalization. Sustainable autonomy and economic efficiency, however, also require some equalization of opportunity so that local councils have an approximately “level playing field” with which to offer services to their residents. This observation points to the crucial importance of equalization transfers in fiscal decentralization design. In the design of intergovernmental transfers, the right order of policy formulation is to first ask and answer the question about which of many possible objectives, besides equalization, the intergovernmental transfer system is to accomplish. The major issue therefore is that different types of transfers have different objectives, and it is important to match grant design according to these objectives. From a Chilean perspective, the two most important lessons to be applied here are: first, to use a separate transfer instrument for each objective pursued by the central government so that its effectiveness can be monitored, and second, that equalizing expenditure needs and the fiscal capacity to provide for those needs across the regions in the country will not only be an equitable outcome that citizens demand, but it will be also a way to reduce inefficient migration in the country to its capital Santiago, simply because people are just looking for better access to basic public services. 79 • Subnational governments must face a hard budget constraint. If subnational governments are not forced to operate with balanced budgets, they become accustomed to looking to the higher-level governments to cover their shortfalls, they will consciously overspend, and will have little incentive to raise their own revenues. As many countries have learned, it is difficult to break the culture of a soft budget constraint. From a Chilean perspective, given the tight borrowing environment that has existed for local governments in the past, the most important lesson to derive here is that not allowing borrowing or having too little borrowing by the new regional governments can be as inefficient as having too much borrowing and soft budget constraints. The right, somewhat delicate, balance must be found so that the new regions can prudently invest in the public infrastructure necessary for their economic development, and in a manner that fairly distributes the financing costs for that infrastructure among the different generations of resident beneficiaries. • Institution building is a necessary companion of decentralization. A prerequisite for successful decentralization is that local governments possess the administrative infrastructure required to effectively carry out the responsibilities assigned to them, regarding the number and quality of staff, and equipment. Administrative capacity is part of the necessary institutions, but there are others that are equally, or more, important. Democratic representation, sound budget processes and institutional arrangements to ensure coordination and cooperation between different levels of government—both at the political and the technical level—are crucial for a multi-tier system of government to work efficiently. This reflects the fact that even if the structure of intergovernmental fiscal relations is very well designed, there will be a need to address the technical, administrative, and political issues that invariably arise in any system. From a Chilean perspective, it seems that the requisite political democratic institutions would be guaranteed. However, it is not so clear that the administrative capacity will be in place without a coordinated effort for training administrative personnel in their new roles and capacities. This element can make or break the success of the new regional decentralization in Chile. 80 Bibliography Bahl, R. and Martinez-Vazquez, J. (2006) “Sequencing Fiscal Decentralization” World Bank Policy Research Working Paper No. 3914, April. Baltaci, Mustafa; Yilmaz, Serdar. 2006. “Keeping an eye on subnational governments: internal control and audit at local levels” (English). WBI working paper series. Washington, DC: World Bank. http://documents.worldbank.org/curated/en/458411468178144890/Keeping-an-eye-on-subnational- governments-internal-control-and-audit-at-local-levels Bergvall, D., C. Charbit, D. Kraan, and O. Merk (2006). “Intergovernmental grants and decentralized public spending”, Working Papers on Fiscal Federalism, No.3. Blöchliger, H. and Vammalle, C. (2009). ‘Intergovernmental grants in OECD countries: Trends and some policy issues’ in General Grants versus Earmarked Grants Theory and Practice: The Copenhagen Workshop 2009. Junghun Kim, Jørgen Lotz and Niels Jørgen Mau (editors). Albertslund: Korea Institute of Public Finance and Danish Ministry of Interior and Health, pp. 126–48. Blom-Hansen, J.(2009). ‘The fiscal federalism theory of grants: Some reflections from political science’ in General Grants versus Earmarked Grants Theory and Practice: The Copenhagen Workshop 2009. Junghun Kim, Jørgen Lotz and Niels Jørgen Mau (editors). Albertslund: Korea Institute of Public Finance and Danish Ministry of Interior and Health, pp. 167–90. Borge, L.E. and Lilleschulstad, G. (2009). ‘General grants and earmarked grants in Norway’ in General Grants versus Earmarked Grants Theory and Practice: The Copenhagen Workshop 2009. Junghun Kim, Jørgen Lotz and Niels Jørgen Mau (editors). Albertslund: Korea Institute of Public Finance and Danish Ministry of Interior and Health, pp. 191–217. Brosio, G. and Piperno, S. (2009). ‘Conditional intergovernmental transfers in Italy after the constitutional reform of 2001’ in General Grants versus Earmarked Grants Theory and Practice: The Copenhagen Workshop 2009. Junghun Kim, Jørgen Lotz and Niels Jørgen Mau (editors). Albertslund: Korea Institute of Public Finance and Danish Ministry of Interior and Health, pp. 218–38. Davey, K. (1988). “Municipal Development Funds and Intermediaries,” Background Paper for the 1988 World Development Report, The World Bank, July. Ebel, R.D. and Taliercio, R. (2005). “Subnational Tax Policy and Administration in Developing Economies,” Tax Notes International, March 7, p. 919-936. Frank, J. and Martinez-Vazquez, J. (2015). Decentralization and Infrastructure: From Gaps to Solutions edited with Jonas Frank, Routledge. Giugale, Marcelo, Vinh Nguyen, Fernando Rojas, and Steven B. Webb. 2000. “Overview.” In Marcelo M. Giugale. and Steven B. Webb, eds. Achievements and Challenges of Fiscal Decentralization: Lessons from Mexico. Washington, D.C.: The World Bank. Hermansson, A. (2009). ‘Specific and general grants in Sweden: What has happened after the grant reform in the 1990’s?’ in General Grants versus Earmarked Grants Theory and Practice: The Copenhagen Workshop 2009. Junghun Kim, Jørgen Lotz and Niels Jørgen Mau (editors). Albertslund: Korea Institute of Public Finance and Danish Ministry of Interior and Health, pp. 239–62. Hirschman, A.O. (1970). Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States. Cambridge, MA: Harvard University Press. ISBN 0-674-27660-4 Hooghe, L., Marks, G., Schakel, A. H., Chapman, S., Niedzwiecki, S., and Shair-Rosenfield, S. (2016). A Postfunctionalist Theory of Governance. Volume I: Measuring Regional Authority. Oxford: Oxford University Press. Kim, J. (2009a). ‘Introduction’ in General Grants versus Earmarked Grants Theory and Practice: The Copenhagen Workshop 2009. Junghun Kim, Jørgen Lotz and Niels Jørgen Mau (editors). Albertslund: Korea Institute of Public Finance and Danish Ministry of Interior and Health, pp. 13–39. Kim, J.(2009b.) ‘General grants vs. earmarked grants: Does practice meet theory?’ in General Grants versus Earmarked Grants Theory and Practice: The Copenhagen Workshop 2009. Junghun Kim, Jørgen Lotz and Niels Jørgen Mau (editors). Albertslund: Korea Institute of Public Finance and Danish Ministry of Interior and Health, pp. 149–66. Lago-Peñas, S. and Martinez-Vazquez, J. (2013) The Challenges of Government Size. Edward Elgar Publishing Limited. Cheltenham, UK. Lotz, J. (2009). ‘Member states' practices for the funding of new competences of local authorities’. Unpublished manuscript. European Committee on Local and Regional Democracy, Council of Europe. Martinez-Vazquez, Jorge. 1999. The Assignment of Expenditure Responsibilities. Paper prepared for the course on Intergovernmental Relations and Local Financial Management, World Bank Institute. Washington: World Bank. Martinez-Vazquez, J. (2000), ‘An introduction to international practices and best principles in the design of capital transfers’. Mimeo. Georgia State University, Atlanta. Martinez-Vazquez, J. (2015). “Tax Assignments at the Regional and Local Levels,” in E. Ahmad and G. Brosio (eds.) Handbook of Fiscal Federalism and Multilevel Finance, Cheltenham, Edward Elgar, UK Northampton, MA, USA. Martinez-Vazquez, J. (2019). “Emerging trends in fiscal transfer systems in selected federations: Implications for India”, chapter 14 in Farah Zahir and Serdar Yilmaz (eds.) Intergovernmental Transfers in Federations, Cheltenham, Edward Elgar, UK Northampton, MA, USA. Martinez-Vazquez, J. and A. Timofeev (2007). “Reforming the assignment of functional responsibilities: principles and a practical guide.” International Studies Program. Andrew Young School of Policy Studies. Georgia State University (September). Martinez-Vazquez, J. and Vaillancourt, F. (2010). Fiscal Decentralization in Developing Countries: Global Perspectives on the Obstacles to Political and Economic Devolution, Cheltenham, Edward Elgar, UK Northampton, MA, USA. Martinez-Vazquez, J., Lago-Peñas, S., and Sacchi, A. (2017), "The Impact of Fiscal Decentralization: A Survey" with S. Lago and A. Sachi, Journal of Economic Surveys, Volume 31, Issue 4, September; pp 1095–1129. 82 Martinez-Vazquez, J., Vulovic, V. (2017) “How well do subnational borrowing regulations work?” in Naoyuki Yoshino and Peter J. Morgan eds., Central and Local Government Relations in Asia, Cheltenham, Edward Elgar, UK Northampton, MA, USA. 2017. Martinez-Vazquez, J. and Civelek, Y. (2019). "Subnational Government Debt Governance: Lessons from Non-Asian Emerging Economies," International Center for Public Policy Working Paper Series WP 1907, Andrew Young School, Georgia State University. McLure, Charles, and Jorge Martinez-Vazquez. 2000. The Assignment of Revenues and Expenditures in Intergovernmental Fiscal Relations. Paper prepared for the core course on Intergovernmental Relations and Local Financial Management, World Bank Institute. Washington: World Bank. Mueller, D. (2006) “Federalism: A Constitutional Perspective,” in Congleton, R.D. and Swedenborg, B. (eds.) “Democratic Constitutional Design and Public Policy,” p: 205-228, MIT Press, Cambridge, MA. Musgrave, R. (1959) The Theory of Public Finance: A Study in Public Economy. New York: McGraw- Hill. Oates, W. (1972) Fiscal Federalism. New York: Harcourt Brace Jovanovich. Searle, R. and Martinez-Vazquez, J. (2006). “The Design of Conditional Grants” in J. Martinez-Vazquez and R. Searle (eds.) The Challenges in the Design of Fiscal Equalization and Intergovernmental Transfers, Springer Verlag, 2006. Shah, A. (1991) “Perspectives on the Design of Intergovernmental Fiscal Relations,” World Bank Working Paper. Washington, DC. Shah, A. (2007a). “Local Budgeting”. Washington, DC: World Bank. © World Bank. https://openknowledge.worldbank.org/handle/10986/6669 License: CC BY 3.0 IGO.” Shah, A. (2007b). "Budgeting and Budgetary Institutions," World Bank Publications, The World Bank, number 6667, September. Sharma, C. (2018), Emerging Dimensions of Decentralization Debate in the Age of Glocalization, Kurukshetra University, Kurukshetra, Haryana, INDIA, http://mpra.ub.unimuenchen. de/6734/. Smart, M., Bird. R. (2009). ‘Earmarked grants and accountability in government’ in General Grants versus Earmarked Grants Theory and Practice: The Copenhagen Workshop 2009. Junghun Kim, Jørgen Lotz and Niels Jørgen Mau (editors). Albertslund: Korea Institute of Public Finance and Danish Ministry of Interior and Health, pp. 40–73. Sutherland, D., Price, R., Joumard, I., (2005). Fiscal Rules for Sub-Central Governments: Design and Impact, Working Paper No.1, OECD, Network on Fiscal Relations Across Levels of Government, COM/CTPA/ECO/GOV/WP(2006)/1 Ter-Minassian, T. ed. (1997). Fiscal federalism in theory and practice. International Monetary Fund. Washington D.C. 83 Ter-Minassian, T. Craig, J. (1997). Control of subnational government borrowing. in Fiscal Federalism in theory and practice, ed. Teresa Ter-Minassian. International Monetary Fund. Washington D.C. Tiebout, C. M. (1956). A pure theory of local expenditures. The Journal of Political Economy 64: 416- 424. OECD/UCLG (2019). Report of the World Observatory on Subnational Government Finance and Investment – Key Findings. Paris. OECD (2019), Making Decentralization Work: A Handbook for Policy-Makers, OECD Multi-level Governance Studies, OECD Publishing, Paris, https://doi.org/10.1787/g2g9faa7-en. Zeikate, S (2002). Investment Transfers; Survey of Ten Developed Countries. Mimeo. World Bank (2004) “World Development Report of 2004: Making Services Work for Poor People” Washington DC World Bank (2008) “Local Government Discretion and Accountability: A Local Governance Framework”. Economic and Sector Work Report No. 40153. Washington, D.C. World Bank (2009) “Local Government Discretion and Accountability: Application of a Local Governance Framework” Economic and Sector Work Report no. 49059-GLB, Social Development Department, World Bank, Washington, DC. World Bank (2018): Program (Scheme) and Project Appraisal Stage Review as part of World Bank’s technical assistance program “Support to the Government of India for Development of Public Financial Reform Strategy (P152996)”. 84 Annexes Annex 1. Practical steps in expenditure assignments STEP 1: Produce a complete inventory of current expenditure assignments, identifying the areas where there is lack of clarity: In many cases this can be a quite a significant undertaking. While in some countries there may be currently no formal assignment and, in some others, the formal assignment of competences may be just a perfunctory list of areas of activity, repetitive and unclear, in practically all countries there will be a de facto assignment that is quite complete, full of ambiguities and disputes in some cases and not so much in others. Current assignments are likely to be scattered across dozens of pieces of legislation and government acts regulating different government services. The point of this first step is to know as well as possible what, in practice, expenditure assignments are and to identify all ambiguities. Without knowing what we have it will be difficult to know where it is that we should go. STEP 2: Determine goods and services that markets will fail to deliver and which, therefore, must be provided within the public sector: This step, which deals with the proper demarcation of the public sector vis-à-vis the private sector, can be a difficult and controversial one, but nevertheless is clearly much needed. Many countries have undergone this process of privatization and decentralization reform over the past several decades. STEP 3: For each non-private good or service, determine the benefit area of the service when provided at the optimal scale: For some services the benefit area is determined by physical or geographical factors. For example, regardless of the boundaries of a jurisdiction that we make responsible for water and sewage, the costs and benefits of this service will affect everybody sharing the watershed and drainage basin. For other goods with spatial distribution of benefits, the benefit area depends on the chosen scale of production. In determining the optimal scale of production and corresponding benefit area, it is necessary to consider long- term trends such as overall pattern of growth. It is also necessary to keep in mind that the “optimal scale” is determined not only by considerations of production efficiency but also the efficiency of the political process stemming from opportunities for participation, civic oversight and overall accountability of public officials. This process will require consultations with and the participation of experts in line ministries and provincial departments and agencies STEP 4: Determine whether “special circumstances” may be present in the delivery of th e service, including: ▪ Spillovers beyond the primary benefit area ▪ National Objectives (“Merit Goods”) such as equal access or special consideration to be given to minority groups, and so on. ▪ Significant income distribution concerns ▪ Cultural/ethnic differences with respect to demand for the service This step will require a good deal of judgment and many consultations. The identification of any of these circumstances requires a national perspective for the most part, and therefore the “Reform Commission” will need to mobilize the expertise of central government agencies and also be guided by current 85 government objectives for redistribution and ethnic issues as reflected in sectoral laws and other legal and programmatic documents. STEP 5: If the function does not involve the aforementioned special circumstances, assign the entire (exclusive) responsibility to the lowest level of government that encompasses the entire benefit area (applying the “subsidiarity principle”). If the function is decentralized too low, then some benefits would spill over, disregarded by that jurisdiction and as a result the service would be underprovided. If the responsible jurisdiction is much larger than the benefit area, then allocation might be inefficient because the government is not close enough to take into account all local specific needs and preferences. Essentially centralization to the next higher level would require the further aggregation of local information used by decision-makers at that level. The loss of information would inevitably lead to less efficient decisions. Size variations among jurisdictions of the same level accompanied by differences in administrative capacity may require –at least temporarily-- asymmetric assignments. STEP 6: Where the conditions are present assign exclusive responsibilities to higher levels of government and where the conditions require it, proceed to assign concurrent or shared responsibilities to two or more levels of government: But, where concurrent assignments are necessary or desirable, it is necessary identify the attributes or sub- functions associated with the particular competence o (regulation, financing, administration, and delivery) and clearly assign each to particular government levels. When any of these attributes or sub-functions is still shared by two or more levels of government, further subdivide that attribute in sub-functions until the latter is exclusively assigned to one level of government. An example is provided in Table VV for education services. Table A1.1. Assigning functional attributes for education Function Setting policies & Financing Provision regulations & production Pre-school education LG(/CG) LG/P LG/P Primary education CG(/RG) CG(/RG)/LG/P LG/P Higher (university) education CG(/RG) CG(/RG)/P CG(/RG)/P Note: “CG” for central government; “RG” for regional government; “LG” for local government; “P” for private. More detailed subdivision of a competence into attributes or sub-functions are typically necessary. For example, Table BB reproduces the division of attributes for primary education used by the OECD (2019): Table A1.2. Key Responsibilities for Education in OECD Documentation Organization of School attended by student. Instruction Instruction time. Choice of textbooks. Curriculum content. Teaching methods. 86 Personnel Management Hire and fire school director. Recruit and hire teachers. Set or augment teacher pay scale. Assign teaching responsibilities. Determine provision of in-service training. Create or close a school. Planning and Structures Selection of programs offered in a school. Definition of course content. Set examinations to monitor school performance. Develop school improvement plan. Resources Allocate personnel budget. Allocate non-personnel budget. Allocate resources for in-service teacher training. STEP 8: If expenditure assignment reform is to bring changes relative to the status quo, it will be necessary to work out beforehand procedures for the necessary transfers of personnel and assets: A new assignment of functional competences is likely to involve the transfer, sometimes massive, of personnel and assets. For example, in the relatively recent “big bang” decentralization in Indonesia, close to two million central government employees and many thousands of service structures were transferred to sub-national government in a very short period of time. In the case of Indonesia, the transfer of assets and personnel was significantly facilitated by the extensive geographical deconcentration of central government services in the years prior to the decentralization reforms in 2001. The transfer of personnel without the transfer of assets and structures can affect the quality and intensity of the decentralization process. These issues bogged down the decentralization process for yeas in Russia and other former Soviet Union countries and still are a problem in some Eastern European countries, like Macedonia and Romania. STEP 9: Make regulatory/institutional arrangements for monitoring, evaluation (including performance targets), and retrofitting: A key underlying requirement for successful fiscal decentralization is regular reporting by local government using a uniform structure of budget classification and chart of accounts. In addition, it is important that subnational accounts are regularly and properly audited. The availability of public data on sub-national local budgets and indicators for service delivery can provide several forms of “yardstick competition” across jurisdictions and therefore enhance horizontal accountability of sub-national officials to their residents. Making sub-national budget data widely available can be a significant challenge. However, many countries have done so; for example, it is now possible in Tanzania to access all local government data on the web. STEP 10: Establish institutions and procedures for review of the assignments, resolving conflicts in the area of delineation of government authority for concurrent responsibilities, and continued exchange of information: Expenditure assignment is never a static structure but rather it will need to adapt to changes and development of new government tasks. Both technological aspects of public services and political circumstances change which affects the optimality of expenditures assignments. Coordination is needed because information generated in the administration and delivery of services is needed for designing 87 regulation and financing this service. There is also a need to provide for explicit mechanisms to manage conflict such as administrative courts or even the Constitutional court, permanent intergovernmental commissions, and parliamentary commissions. Good examples of how these institutions can work are provided in the cases of Germany and Spain. Other countries have introduced this type of institution in the laws but for the most part thy shave never been implemented. 88 Annex 2. Main responsibilities of regional/state government level in Latin America Argentina Bolivia Brazil Colombia Ecuador Mexico Paraguay Peru Uruguay 1. General Administration Administration Passport Supervision of Internal Internal public services of all provincial and operation of issuance municipal affairs administration; administration services; general services; Enhance citizen Coordination Operation and participation and oversight of maintenance of municipal public buildings governments and facilities; Administrative services 2. Public order Fire protection Departmental Administration Risk and Coordination State public and safety (shared); Justice civil security and command of disaster with national order and safety and police plans, programs military police management police (state and (shared) and projects; and fire-fighters; special police) Provide support Civil police for police forces operations, and prisons; administration of the prison system (when delegated by federal law); Road traffic control 3. Economic Economic Plan and execute Agriculture and Rural Promote Roads (state Transportation Regional Transit and affairs/transports promotion; economic and rural development; agricultural roads, rural road development roads; Road Clusters social development; Regional and other development, planning; Road management; development; development Regional policies; productive maintenance of infrastructure, Departmental Promotion and plans; Rural economic Regional activities; federal energy, and urban financing of electrification development and territorial Road networks secondary communications; transport; economic and energy support to local planning; and facilities roads); Regional Promotion of Promote the services; provision; enterprises and Traffic in non-urban transport and employment; development Electricity; Employment and entrepreneurship; management areas transit; Some Tourism; of agriculture, Interprovincial tourism Regional tourism; airports; Support to livestock, roads; Public promotion, Ports; Agriculture, economic and industry and transport infrastructure Management and rural productive trade (shared) building and maintenance of development activities and delivery of state highways and tourism SMEs services for and roads; (shared); production; Management of Economic Building and inter-municipal affairs and management of transport industrial departmental policies (shared) train lines, road routes and airports 89 Argentina Bolivia Brazil Colombia Ecuador Mexico Paraguay Peru Uruguay 4. Natural Natural heritage Environmental Environmental Environmental Protection of Preservation Promote Protection of Environmental resources conservation; licensing protection management; environment of the sustainable use the protection management; Integrated waste (shared); Watershed and (shared); environment of forestry environment Sewage management Environmental micro-basin National parks and natural resources and and (shared) (responsibility education management; (shared); Own resources biodiversity; sustainable shared with the (shared); Hydraulic state Preservation and development national and Protection, dredging and environmental management of of natural municipal preservation and river cleaning; standards natural reserves resources; levels) restoration of Forestry and and protected Waste regional natural reforestation; natural areas collection resources Irrigation systems management 5. Housing and Water Departmental Supply and Coordination Housing Spatial planning; Public works; Land-use Construction community management land use treatment of water and co- (management); Water Urban planning and housing amenities (shared) planning and basic financing of Urban and management and planning (part of its sanitation; Energy water schemes land use co-financing of execution and production (when planning (in water regulation); delegated by coordination infrastructure Territorial federal with other (shared) and urban government); levels); Plans planning; Metropolitan land and programs Street use planning; for housing in cleaning; Housing rural areas Public construction and lighting; settlement Sanitation; upgrading Cemeteries (shared) 6. Health Healthcare Infrastructure Hospital services; Public health; Healthcare Health Participation in Public (shared) building and Preventive Services for (shared): promotion the management hygiene and provision of healthcare; the uninsured organization and of public health health equipment and Secondary and poor operation of medical supplies tertiary health population; healthcare for hospitals services Operation of services for the the hospital uninsured network population; Primary care for the rural and urban poor; Health services; Administration and maintenance of hospitals for primary care; Preventive and reproductive care 90 Argentina Bolivia Brazil Colombia Ecuador Mexico Paraguay Peru Uruguay 7. Recreation, Sports; Culture; Regional Cultural, arts, Culture (public Access to Libraries; culture & Oversight of museums; sports and libraries) and culture; Museums; religion artistic and Libraries, recreational recreation Strengthening of Exhibitions; architectural Cultural activities activities for regional artistic Nurseries; patrimony and preservation the rural areas and cultural Theater; of heritage (in institutions Music; coordination Sports; with the Gardens; parishes) Zoos; Planetariums 8. Education Secondary School Secondary and Management Primary and Education Management of education, infrastructure higher education; of teacher and secondary infrastructure preschool, including building; Basic Support to administrative education primary, vocational services and municipal personnel in (shared); State secondary, and education and provision of administration of basic and universities; higher non- technical educational basic education primary Adult education university training; material education programs; education Tertiary Indigenous and education special (shared) education; School-lunch programs 9. Social Unemployment Implementations Social Protection; Equity and Poverty School meals Activities and protection subsidies of social projects Support inclusion; alleviation policies for (shared) for childhood, municipal Citizen (shared); Social specific adolescence, provision of care participation; protection population women, the services Protection of (shared) groups elderly and children and including food (women, people with youth assistance for children, disabilities. the poor young people, the elderly and people with disabilities) Source: Own elaboration based on 2019 Report World Observatory on Subnational Government Finance and Investment, Country Profiles 91 Annex 3. Salient issues in the political economy of decentralization Political economy issues are pervasive among the reasons why country may want to decentralize, as well as in the final architecture of decentralization that gets adopted. As already emphasized in different part of the report, the reasons why countries attempt to decentralize are often political, even though experts tend to justify centralization on the basis of objective goals, such as the overall increase in the efficient allocation of public resources, increasing public access to basic services or strengthening democratic representation and government accountability to citizens. However, countries decentralize at time because of political opportunity considerations, like one a political party realizing that it may get more electoral support in the national periphery than in the core capital are of the country (as was the case in Peru, prior to the 2001 decentralization reform). Political economy considerations tend to affect for example the vertical organization of government: how many tiers of government and how many government units in each tier. For example, considerable diversity at the regional level based on ethnic-linguistic or even econ0omic differences puts political pressure on system to decentralize at the regional level and not only at the local level, and also to consider the introduction of asymmetric systems by providing more decentralized authority to some regions than to others (as has been the case of Indonesia with the provinces of Aceh and West Papua; of with the Basque Country and Navarre in Spain). Political economy considerations may also be behind the seemingly difficulty many countries appear to have with introducing clear and transparent assignment of functional expenditure responsibilities. Having an unclear assignment may allow central authorities to impose spending mandates on subnational units without providing the necessary funding or give a sense that they can intervene or blame subnational government any time services are not well provided and citizens complaint. Similarly, Subnational government may be happy to live with unclear assignments because it gives a greater capacity or flexibility in the allocation of scarce resources or the ability to blame central authorities for any actual deficiencies. Similar perverse incentives, as highlighted in the report, may work at the central and subnational levels in devising a decentralized system that grants little revenue autonomy to subnational governments. While central authorities may want to keep revenues centralized to retain power and exert influence on the behavior of subnational units, these latter may be happy not to have to collect their own revenues because that will bring more oversight and accountability to local residents, while transfers dependence from the center does not carry political costs and can be used to blame central authorities or any shortcomings in public services. Many aspects of the design of a system of intergovernmental transfers can be also subject to political economy consideration and pressures. For example, both central and subnational authorities may favor discretional and political motivated transfers over objective and formula-driven transfers because they feel they can get ahead of the game and support their own political designs. Even in the case of borrowing and subnational credit, there are clear moral hazard incentives for subnational government to over borrow and spend and hope for central government rescues and bailouts, while central authorities may like to operate with soft budget constraints for subnational government because that is a way to exercise and extends their political power and influences. 92