Chapter 2. Understanding decentralisation systems

This chapter provides a definition of decentralisation, which covers three interdependent dimensions: political, administrative and fiscal. There can (or should) be no fiscal decentralisation without political and administrative decentralisation. The chapter clarifies the scope of the decentralisation concept, which is often confused with deconcentration for example. Finally, the chapter provides a comprehensive picture of the diversity in territorial organisation and decentralisation systems in OECD countries and beyond.


The statistical data for Israel are supplied by and under the responsibility of the relevant Israeli authorities. The use of such data by the OECD is without prejudice to the status of the Golan Heights, East Jerusalem and Israeli settlements in the West Bank under the terms of international law.

What is decentralisation and how can it be measured? Although widespread and extensively analysed, decentralisation often has different meanings and different scopes. Several definitions exist reflecting different understandings of decentralisation, which adds to an already complex reality. This also reflects the diversity of decentralised governance systems around the world. The forms and extent of decentralisation vary greatly from one country to another – and even within the same country. Large variations exist in terms of responsibilities carried by subnational governments, of local decision-making powers, of resources available to meet the needs, including the ability to raise own-source revenues. There are also varying degrees of upward and downward accountability and varying ranges of central government control. The heterogeneity of experiences on the ground also explains the great difficulty in measuring (and assessing) decentralisation. Fiscal indicators, although useful, remain insufficient to reflect this diversity and have to be complemented by other types of institutional and governance indicators to get the facts right about decentralisation.

A multi-dimensional concept with different motivations

What is decentralisation? (and what it is not)

It is important to clarify the scope of the concept of decentralisation, since it is often confused with other concepts, such as deconcentration, devolution and delegation. The latter two are different degrees of decentralisation (Rondinelli, Nellis and Shabbir Cheema, 1983[1]). Some countries associated the “decentralisation” concept with specific dimensions like delegation or localism, but the full decentralisation concept is broader. Overall, there are no clear-cut frontiers within decentralised governance systems. Rather, there are different degrees of decentralisation, depending on the extent of political, administrative and financial powers that have been transferred to lower levels of government, and on the balance of central-subnational relations.

Decentralisation and devolution

Devolution is a subcategory of the decentralisation concept. It is a stronger form of decentralisation as it consists of the transfer of powers from the central government to lower-level autonomous governments, which are legally constituted as separate levels of government. This was the path chosen by the United Kingdom in 1998, when the three devolved nations of Northern Ireland, Scotland and Wales were established, with a directly-elected “national assembly”/parliament and their own government. Major governing power and responsibilities were transferred to them, creating an asymmetric decentralisation system between devolved nations (with different powers) and England (no regional governments).

Decentralisation and federalisation

The next stage after devolution is federalisation, although some federal countries may be in reality quite centralised systems, with few powers exercised by subnational entities (Box 2.1).

Box 2.1. Decentralisation in federal and unitary countries

Decentralisation processes take different forms depending on the form of the state, i.e. mainly unitary or federal (including confederations), although an intermediate form of “quasi-federalism” exists in some countries. A minority of countries have a federal system of government: of the 193 UN member states, 25 are governed as federal countries (40% of the world population) and 168 are governed as unitary states (Forum of Federations:

Federal countries may not be the most decentralised (“centralised federalism”) and some unitary states may be more decentralised than federal ones.

In addition, in federal countries, the degree of decentralisation may differ within the same country as state constitutions and legal systems for local governments may vary from one state government to another. Local government autonomy can differ greatly depending on their state (e.g. in India).

Federal countries

In federal countries (or federations), sovereignty is shared between the federal government and self-governing regional entities (the federated states), which have their own constitution in most cases (Canada is an exception), parliament and government. In a federation, the self-governing status of the component states may not be altered by a unilateral decision of the federal government.

Powers and responsibilities are assigned to the federal government and the federated states either by provision of a constitution or by judicial interpretation. In general, federal governments have exclusive and listed responsibilities such as foreign policy, defence, money and criminal justice system while federated states have extensive competencies.

In addition, in most federal countries, in particular, older ones, local governments are “creations” of the federated states and fall directly under their jurisdiction (this is not the case in all federations, e.g. Brazil where municipalities are not subordinate to the states in which they are located). Local government status, organisation, responsibilities and financing are defined by state constitutions and laws, and often differ from one state to another. Governed by state legislation, local governments do not have independent relations with the federal government. Local government reforms are decided by the federated states and not the federal power, which has no say on those matters (e.g. Australia, Canada and the United States).

Unitary countries

A unitary state is a state governed as a single power in which the central government is ultimately supreme. The unitary states are “one and indivisible” entities and sovereignty is not shared. This means that citizens are subject to the same single power throughout the national territory.

This does not preclude the existence of subnational governments, also elected directly by the population and with some political and administrative autonomy. But subnational governments exercise only the powers that the central government chooses to delegate or devolve. Unitary states are thus more or less decentralised, depending on the extent of subnational powers, responsibilities and resources, and the degree of autonomy they have over these different elements. In a unitary state, subnational units can be created and abolished and their powers may be broadened and narrowed by the central government.

Some unitary countries recognise autonomous regions and cities, which have more powers than other local governments because of geographical, historical, cultural or linguistic reasons.

Quasi-federal countries

Between these two main forms, there is an intermediate situation, that of “quasi-federal”. This status of “quasi-federation” is however not recognised as such. It applies to unitary countries with federal tendencies, i.e. having some characteristics of a federal country. Autonomous regions have less room to define and reform local government functioning than regions in federations. Basic elements of local government functions and financing are often written out in national constitutions. Even if substantial autonomy is given to autonomous regions in relation to lower tiers through primary and/or secondary legislative powers, it is often a competency that is shared with the central power. This is the case in South Africa and Spain. Spain is constitutionally a unitary state but in reality, a quasi-federation with regions having a large degree of autonomy. Organising the municipalities and provinces and changing municipal boundaries within the regional territory is the exclusive responsibility of the Autonomous Communities, but their functions and finances are decided within the framework of the national law (Article 148 of the constitution).

Sources: OECD (2017[2]), Multi-level Governance Reforms: Overview of OECD Country Experiences,; OECD-UCLG (2016[3]), Subnational Governments around the World: Structure and Finance,

Decentralisation and localism

Decentralisation may be approached in a broader sense such as in England (United Kingdom) with the 2011 decentralisation reform, called “localism”. However, localism should not be confused with decentralisation in its traditional meaning. Being part of the broader “big society” (vs. “big government”) project designed by the British government, the reform consisted in transferring several responsibilities from the state not only to local authorities but also to civil society, such as co-operatives, neighbourhood communities, citizen and volunteer groups, charities, social enterprises, etc., calling for a better “participatory governance” and greater community empowerment.

Decentralisation and delegation

Delegation is a form of decentralisation, but is milder than devolution. It involves transferring some decision-making and administrative authority for well-defined and specific tasks from the central government to semi-autonomous lower-level units, such as state-owned enterprises, or urban and regional development corporations (UNDP, 1999[4]). Functions can be delegated through regulations or contracts. Semi-autonomous entities remain under the indirect control of the central government and their delegated tasks may be withdrawn in a unilateral manner.

Decentralisation and deconcentration

Decentralisation and deconcentration are sometimes used interchangeably. However, they are profoundly different. In the first case, there is a transfer of power from the central government to autonomous/elected subnational governments. In the second case, there is a geographic displacement of power from the central government to units based in regions (territorial administration of the central government, line ministerial departments, territorial agencies, etc.). These deconcentrated state services are part of the national administration and represent the central government at the territorial level. Unlike subnational governments, deconcentrated state services are a hierarchical part of the central government level. They are not legal entities or corporate bodies. This means that deconcentrated territorial bodies do not have their own political leadership and decision-making power. They do not have their own budgets, which are typically included in the national budget. They do not raise revenues, cannot incur liabilities by borrowing on their own or engage in financial transactions and do not have their own assets. Their staff is part of the national civil service (OECD, 2018[5]; 2017[2]; Boex, 2011[6]).

In practice, distinguishing between decentralisation and deconcentration systems is a challenge. In many countries, decentralised and deconcentrated systems co-exist, such as in Estonia, Finland, France, Italy, Poland and Sweden. Decentralisation does not mean that the central government cannot maintain certain functions at the local level or make them evolve and adapt to institutional changes in the governance system. Depending on the country, the role of state representatives at the territorial level in a decentralised system can, however, differ, from a representational function to a more significant role. State territorial representatives may thus be responsible for implementing national policies at the regional and local levels, ensuring that they are in line with subnational government policies. In some countries, state territorial representatives also carry out legal and fiscal oversight functions over local government actions. They may also play a co-ordination role between the different stakeholders, acting as a “pivot” of the administrative system, facilitating multi-level government dialogue on the ground, and sometimes acting as an advisor and “mediator” able to reconcile different perspectives. Finally, deconcentrated state services may also provide national public services at the territorial level.

In France for example, despite the different decentralisation laws, the national government is still very active on the local scene, playing a leading role in many areas. It has maintained, at both regional and departmental levels, a strong and powerful prefectural administration led by a prefect (préfet), as well as local directorates of various ministries placed under their authority, so-called “deconcentrated services”. According to the constitution, the prefect is the direct representative of the prime minister and every minister at the departmental level, implementing government policies and their planning. The prefect is responsible for national interests, administrative supervision and compliance with laws and is in charge of public order (OECD, 2017[2]).

In some countries, the co-existence of decentralised and deconcentrated systems may be even more accentuated, resulting in an intricate system of governance. This can be very confusing, generate tension and reduce transparency and accountability. In some countries, subnational governments are “mixed” or “dual” entities (i.e. both deconcentrated and decentralised), where the same subnational body is both a deconcentrated administration representing the central government (the executive power) and an elected autonomous self-government (the deliberative power). This is the case in Turkey and Ukraine (OECD, 2018[5]).

So, what is decentralisation?

Decentralisation refers to the transfer of powers and responsibilities from the central government level to elected authorities at the subnational level (regional governments, municipalities, etc.), having some degree of autonomy. Subnational governments are a legal entity directly elected by universal suffrage and having their own budget, staff and decision power. Both devolution and delegation of tasks fulfil this definition, but to different degrees.

Box 2.2. Defining decentralisation: The OECD approach

The definition of decentralisation used at the OECD is the following: “decentralisation consists in the transfer of a range of powers, responsibilities and resources from central government to subnational governments, defined as legal entities elected by universal suffrage and having some degree of autonomy”. Subnational governments are thus governed by political bodies (deliberative assemblies and executive bodies) and have their own assets and administrative staff. They can raise own-source revenues, such as taxes, fees and user charges and they manage their own budget. Subnational governments have a certain decision-making power, in particular, they have the right to enact and enforce general or specific resolutions and ordinances.

In this definition, decentralisation is not only about transferring powers, responsibilities and resources. It is also about reconfiguring the relationships between the central government and subnational governments towards more co-operation and co-ordination. Managing “mutual dependence” requires a profound change of structure, practice and culture within the central government itself, which represents substantial challenges for central governments around the world (Devas and Delay, 2006[7]).

Overall, this definition is associated with the concept of Local Self-Government as defined in the European Charter of Local Self Government (1985), in UN-Habitat’s International Guidelines on Decentralisation and Strengthening of Local Authorities (2007) and, more recently, in the African Charter on the Values and Principles of Decentralisation, Local Governance and Local Development (2014).

Sources: OECD (2018[5]), Practical Methodological Guide for the World Observatory on Subnational Government Finance and Investment, OECD, Paris; Boex, J. (2011[6]), Exploring the Measurement and Effectiveness of the Local Public Sector: Toward a Classification of Local Public Sector Finances and a Comparison of Devolved and Deconcentrated Finances.

It is important to note that decentralisation may also be territorial or functional. Subnational governments are not always defined by their territory, i.e. the geographic boundaries within which they exercise their authority and carry out their responsibilities. While they are elected, subnational governments can also be defined just by the specific functions they perform. Examples of these “special-purpose subnational governments” are school boards, transport districts, water boards and sanitation districts. This is also known as “functional decentralisation” as opposed to “territorial decentralisation”. In the United States, for example, the Census Bureau recognises five basic types of local government, including two special purpose governments: school districts and “special district governments”. Legislative provisions for school district and special district governments are diverse and the basic pattern of these entities varies widely from state to state (OECD, 2018[5]; Department of Commerce (U.S. Census Bureau), 2012[8]). In Korea, the local government system comprises 17 educational offices, which are independent elected entities. In the Netherlands, Regional Water Authorities (RWAs) are subnational government bodies and their autonomous tasks are defined in the Regional Water Authorities Act (Waterschapswet).

Figure 2.1. Defining decentralisation in unitary and federal countries

A multidimensional process

The concept of decentralisation covers three dimensions: political, administrative and fiscal. These dimensions are inter-dependent: there can (or should) be no fiscal decentralisation without political and administrative decentralisation. On the other hand, without fiscal decentralisation, political and administrative decentralisation are meaningless.

  • Political decentralisation sets the legal basis of decentralisation. It involves a new distribution of powers according to the subsidiarity principle, between different tiers of government, with different objectives, and often with the aim of strengthening democracy. Thus, it refers to the way in which subnational administrators are selected – i.e. by appointment or by election.

  • Administrative decentralisation involves a reorganisation and clear assignment of tasks and functions between territorial levels in order to improve the effectiveness, efficiency and transparency of national territorial administration. It generally relates to the transfer of planning, financing and management decisions on some public functions to lower levels of government.

  • Fiscal decentralisation involves delegating taxing and spending responsibilities to subnational tiers of government. In this case, the degree of decentralisation depends on both the amount of resources delegated and the autonomy in managing such resources. For instance, autonomy is greater if local governments can decide on tax bases, tax rates and spending allocations.

In theory, these three dimensions – the distribution of powers, responsibilities and resources – are complementary and closely interconnected (Figure 2.2). Links should be carefully considered to maximise the chance of success (Chapter 5). In practice, finding the right balance between these dimensions, and finding the right sequencing, represent major challenges to making the most of decentralisation reform. There are often one or two missing links and bad sequencing. For example, the political dimension is often insufficiently considered in some decentralisation processes, with little democratic legitimacy of subnational governments, no real accountability mechanisms or weak citizen involvement at the local level.

Figure 2.2. Political, administrative and fiscal decentralisation

Motivations for decentralisation

Decentralisation reforms are and have been implemented for a wide variety of reasons. According to Ivanayna and Shah, “hugely complex factors such as political transition in Eastern Europe, the end of colonialism, the globalisation and information revolution, assertion of basic rights of citizens by courts, divisive politics and citizens’ dissatisfaction with governance and their quest for responsive and accountable governance have been some of the contributing factors in gathering this storm” (2014[9]). As Hooghe et al. (2016[10]) point out, decentralisation trends are also the result of unravelling the excessive centralisation brought about by authoritarianism, nationalist state building and wars over the course of the late 20th century.

Decentralisation has also been implemented as part of state reforms to improve efficiency and quality of public services, to enhance regional and local productivity and growth, to meet fiscal consolidation objectives in the aftermath of the recent economic crisis, or in response to the institutional programmes of supranational organisations. Sometimes the motivations are not positive, as a number of countries have decentralised expenditures to shift deficit downwards. Historically, highly decentralised states were often the result of compromises required in countries with a diverse population, in part to ensure political stability.

The rising decentralisation paradigm has been driven by three main categories of factors: political drivers, economic drivers and megatrends.

Decentralisation is primarily a political choice

Engaging in a decentralising reform is ultimately a political issue and thus should be conceived and pursued as part of a larger political reform process, including, for example, territorial development reforms, reforms of the judiciary, civil service and regulatory frameworks, while also building greater accountability and reducing political instability. In theory, decentralisation reforms promote a new conception of governance no longer based on hierarchical, top-down and vertical relations but adding a bottom-up and co-operative perspective, with more transparency, accountability and participation, in particular by citizens. The principles of local democracy and subsidiarity defined as the “precept […] that public policy and its implementation should be assigned to the lowest level of government with the capacity to achieve the objectives” form the fundamental basis of this political choice (Oates, 1999[11]). Decentralisation 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 (Box 2.3).

In theory, democratisation is the basis for genuine decentralisation as decentralisation involves local elections based on pluralism and greater participation of citizens. In practice, as noted above, local democracy may be neglected as well as mechanisms allowing real involvement of the population in local affairs. Restoring or building democratic systems has been a strong incentive for decentralisation in several Asian, Latin American and Central and Eastern European countries. In Korea, decentralisation started in 1987 with the “Declaration for Democratisation” and gained momentum in 1988 with the reform of the Local Autonomy Act, followed by the organisation of the first local elections in 1991 and 1995.

Several moves towards decentralisation around the world have been motivated by this quest for more local democratic control. In Asia, achieving more democratic political outcomes has also been a strong incentive for decentralisation in several countries such as i.e. India, Indonesia or Korea with pro-democratic movements and popular mobilisation rejecting centralised autocratic governments and dictatorships. In Japan, the push for decentralisation started during the post-World War II period and was viewed as a means of achieving more democratic political outcomes. The promotion of a democratic system of local government was part of the national agenda (Chatry and Vincent, forthcoming[12]).

The same occurred in Central and Eastern European countries in the early 1990s as a reaction to the failures of the communist centralised state over the previous four decades and in several Latin American countries where the return to democratic government was associated with decentralisation. In Africa, South Africa’s decentralisation process was an essential component of its transition from apartheid to democracy (OECD, 2008[13]; 2011[14]).

Decentralisation has also been a way to ensure greater political stability by preventing the disintegration of the “nation-state” in countries with strong regional identities or in reconstructing countries afflicted by conflicts, such as Ukraine (OECD, 2018[15]).

Box 2.3. Decentralisation, a constitutional principle in several countries

Decentralisation is enshrined in several national constitutions, as a fundamental principle to deepen democracy, improve the balance of powers between higher and lower levels of government and ensure the effective participation of citizens in decision-making. In theory, this constitutional status protects subnational governments’ democratic structure and functioning from central government interference or dismissal. The constitution thus represented a safeguard against arbitrary dismissal of local government (Shah, 2017[16]).

In Peru for example, the current decentralisation process started in 2002, when congress, driven by democratic and economic objectives, constitutionally declared Peru a “decentralised state” (Article 43 of the constitution). In Colombia, the adoption of the new constitution in 1991 significantly enhanced the process of decentralisation, declaring that “Colombia is a legally organised social state under the form of a unitary, decentralised Republic with autonomous regional entities” (OECD, 2014[17]; 2016[18]).

In federal countries, the federal constitution is the key to defining the rules that determine power sharing between the federal government and state governments, including fiscal arrangements. However, the federal constitution does not always take into consideration the principle of decentralisation at the local government level. In Australia for example, municipalities are not explicitly recognised by the Commonwealth Constitution, despite failed or abandoned referendums proposing constitutional recognition in 1974, 1988 and 2013. In some federations, however, the guaranteed principle of local self-government is set out in the federal constitution, e.g. Austria, Belgium, Germany, Mexico and Switzerland.

Economic motivations are also important drivers of decentralisation

The economic approach to decentralisation emphasises the improvement of local public services. The idea is that local governments have better information regarding local spending needs and preferences, and hence may better satisfy certain needs of the population, at a lower cost, than the central administration (Chapter 4).

The motivations for decentralisation may be 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), to increase equity in access and services and tailor policies to local contexts and population needs. These have been important motivations for the support of decentralisation by multilateral organisations, such as the International Monetary Fund (IMF) and the World Bank.

Mega-trends and decentralisation

Mega-trends linked to digitalisation, the globalisation of economic activity and urbanisation, and contribute to the stronger role of subnational governments. Globalisation implies that free movement of capital between and within countries also means that cities can compete for foreign investment, a task once monopolised by central governments (Olowu, 2003[19]). On the other hand, globalisation 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-emphasised in a context of backlash against globalisation. In this context of democratic crisis in many countries, the role of local governments has also been emphasised 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”. The relationship between the global and the local is often referred to as “glocalisation” (Sharma, 2018[20]).

Urbanisation is another major global trend that affects decentralisation. Today, more than 50% of the world’s population lives in cities. This figure is projected to reach 85% by 2100. Within 150 years, the urban population will have increased from less than 1 billion in 1950 to 9 billion by 2100. This period of rapid urbanisation will also have experienced the rise of the megacity, which is defined as a metropolitan area with a total population in excess of 10 million people. In 1950, New York and Tokyo were the only megacities, but by 2014 their number had increased to 28 (OECD, 2015[21]). Cities are growing in terms of population, infrastructure and networks. They are becoming more and more complex.

The information revolution and digitalisation have weakened the ability of governments to control information flows. Information and communication technology (ICT) and other technological changes (blockchain, geographic information system (GIS), robotics) have also multiplied the opportunities for local governments to improve the ways they communicate and involve citizens (e-democracy and ICT-based participation, accountability and transparency in local governance). Digital tools also help governments provide local public services (e-government), manage public resources in a more efficient manner (e.g. for tax collection), improve staff capacity and management and adopt new public management models. ICT can improve the relationships between the central and subnational governments, facilitating the shift towards more decentralised governance practices.

Figure 2.3. Drivers for decentralisation

Different paths to decentralisation

Big bang vs. incremental approaches

Decentralisation reforms can be carried out in different ways: with a “big bang” approach, or step-by-step in an incremental manner or in waves (e.g. Japan with First and Second Decentralisation Promotion Reforms or France with the three “Acts on Decentralisation” from 1982 to 2015).

A “big bang” approach has two defining characteristics: i) it entails a holistic (comprehensive) and integrated approach of the three main components of decentralisation (political, administrative and fiscal); ii) it is rapidly implemented over a short period. One of the main objectives of this approach is to ensure that major measures to implement political, administrative and fiscal decentralisation will be adopted as a single package, without missing components, and that “all pieces of the puzzle fit together – i.e. the desired balance in autonomy and accountability is achieved while providing incentives for cost efficiency”. “Lightning speed” represents the best use of this window of opportunity, as in all reforms there are losers, who can take advantage of a long period to get organised and build coalitions to circumvent or block the reform (Shah and Thompson, 2004[22]).

However, a “big bang” approach may have some drawbacks. It can be perceived as imposed from the centre too rapidly, and at the risk of being rejected as a transplant that has not been successful. Several countries have had to step back because the decentralisation process was too rapid and strong. This was the case, for example, in Central and East European countries, as well as in many developing countries. Some decentralisation laws have never been implemented, or only partially because the agenda was too ambitious and unrealistic.

An incremental approach has some advantages. Introducing the different measures in sequence can help their implementation, as well as facilitate necessary adjustments. It can also permit a progressive upgrade of subnational government capacities, ensuring that they will be more capable of handling newly assigned responsibilities and effectively managing their resources. A gradual reform process can give more opportunities to convince and engage citizens in the process, to gain support and build greater consensus through consultations, information, public debates, etc. Finally, each step can create an impetus for further reforms from the central government, local authorities and civil society.

Pilot experiments (e.g. the free commune experiments in Denmark, Finland and Sweden) and experimental regionalisation (e.g. in Finland, France and Sweden) represent interesting approaches to decentralisation, as they can demonstrate the effectiveness of reforms and pave the way for further change on a larger scale (OECD, 2017[2]).

Regardless of the path chosen, decentralisation is a never-ending process: modifying the governance system implies continually adapting to the new system. In an already decentralised system, decentralisation is more a process of reconfiguration of the multi-level system. It can take place without major transfers of responsibilities but more adjustments of the current system with some reassignments of functions or fiscal resources such as in France and the Netherlands.

Decentralisation is also a dynamic process. Drivers for reforms may thus vary over time depending on the political, economic, social and budgetary contexts. Some factors may be cyclical depending on economic circumstances. For example, the 2008 global crisis accelerated decentralisation reforms.

Moreover, international experiences show that design and implementation are the most challenging phases of decentralisation reform. Many decentralisation reforms have stalled, failed and been cancelled, postponed or even reversed. Others have been modified and adjusted along the way, or partially implemented. To generate expected benefits, additional and complementary reforms are often needed to correct potential deviations and improve multi-level governance mechanisms.

Top down or bottom up approaches

Decentralisation reform may be carried out in top-down or bottom-up or ways. Historically, bottom-up decentralisation processes have been a dominant mode of decentralisation in North America and Northern Europe. Today, there are few examples of a bottom-up approach. Regional reform in Sweden is one of these examples, with reforms promoted by the top but by leaving the “bottom” to decide. Unlike many countries, the Swedish government did not impose a single regionalisation model, but experimented with “pilot regions” and permitted heterogeneity across regions in terms of governance bodies and regional responsibilities. This approach towards regionalisation has created scope for learning, fine-tuning the reform and fostering consensus (OECD, 2017[23]). A top-down approach, through which the national government decides to shift part of its responsibilities downwards, is more frequent and has been the dominant process of decentralisation followed in Southern Europe and all developing and transition countries (Shah and Thompson, 2004[22]). In these countries, however, there are some cases where decentralisation has been pushed forward by local and regional elites as a strategy for mobilising and maintaining regional power bases (Devas and Delay, 2006[7]).

In the case of the top-down approaches, the reasons to decentralise were far from political or economic (local democracy, accountability, efficiency, etc.) but more focused on short-term considerations (political calculations, shifting responsibility for unpopular measures, moving fiscal burdens to the bottom). This was, for example, the case in some countries during the recent financial crisis, when some social tasks were decentralised without real fiscal compensation, hence forcing local governments to play a “social buffer” role (“decentralisation of the crisis”). Most of the time, this type of approach results in an incomplete decentralisation process, with either the political or the fiscal dimensions missing.

Decentralisation reforms have also been strongly promoted (or even imposed) by the international community (e.g. multilateral banks, international agencies and donors). In 2008-09, the support to Greece from the European Central Bank, the European Commission and the IMF included territorial and institutional reforms (e.g. the Kallikratis reform). The goal was to streamline territorial organisation and to reduce and optimise public spending by reinforcing decentralisation. Decentralisation has also been advocated in Africa and Asia by multilateral banks, international agencies and donors concerned by the failure of central governments to efficiently deliver services and address poverty. Endorsing the economic arguments for decentralisation, sometimes coincided with the liberal agenda of reducing the role of the central government under the banner of “good governance”, but with mixed results, as pre-conditions for effective decentralisation were not met (Devas and Delay, 2006[7]; Shah and Thompson, 2004[22]) (see Chapter 4).

Decentralisation often goes hand in hand with other reforms

Decentralisation is not only a public sector reform. It should be viewed in a comprehensive and “polycentric” way, including the observation of interactions between public entities and private stakeholders (profit or non-profit), in particular citizens and businesses (Ostrom, 2010[24]). Decentralisation reforms are often accompanied by other types of multi-level governance reforms, notably territorial and public management reforms (OECD, 2017[2]). These three categories of reform are often interrelated and complementary (Box 2.4 and Figure 2.4).

A territorial reform can be partly driven by a decentralisation reform. An increasing number of tasks transferred to local governments may put pressure to increase their size and capacity in order to cope with the additional responsibilities. This is the case with the current Ukrainian reform, where municipal amalgamation represents a first step and “a platform for decentralisation” (OECD, 2018[15]). To receive new responsibilities and funding, municipalities have to merge into unified territorial communities.

Box 2.4. Institutional, territorial and management reforms: Three main categories of multi-level governance reforms

Institutional reforms aim at re-organising powers and responsibilities across levels of government. They thus include either decentralisation or recentralisation processes, consisting in the transfer of tasks, assets, human and fiscal means from the central government to subnational governments and vice versa, and redefining relationships across levels of government.

Territorial reforms aim at re-organising territorial structures, often by updating and “re-scaling” regional and local government administrative areas, hence modifying their geographic boundaries. Their goal is to reach a better match between the size of subnational structures and their responsibilities and functions.

Public management reforms aim at re-organising subnational government administrative and executive processes, including human resources management, financial management, e-government, etc.

Figure 2.4. Institutional, territorial and management reforms are closely interrelated

Source: OECD (2017[2]), Multi-level Governance Reforms: Overview of OECD Country Experiences,

The reverse is also true, i.e. a decentralisation reform partly driven by a territorial reform. In France, for example, the 2015 regional reform was first and foremost a reform consisting of amalgamating 22 existing regions into 13 regions to build larger and stronger regions. The rationale behind the reform was to generate savings and reach efficiency gains, build more homogenous regions from a socio-economic point of view and grant them sufficient recognition to engage in international and European co-operation. Eight months later, the law on the “Delimitation of Regions” was followed by another law that modified the allocation of responsibilities across levels of government, including the strengthening of regional responsibilities.

Finally, public management reforms may also be introduced in relation to institutional and/or territorial reforms, either simultaneously or successively, as they provide the opportunity to review and modernise management and governance processes.

Measuring decentralisation is a challenge

Competing definitions of decentralisation, its multi-dimensional nature and the heterogeneity of experiences on the ground explain the great difficulty in actually measuring decentralisation. Fiscal indicators are useful in providing a macroeconomic view of decentralisation but remain partial – only focusing on fiscal aspects – and may lead to a distorted interpretation of reality when they are not complemented by an analysis that goes beyond such indicators. Although crucial, fiscal indicators are insufficient to get the facts right about decentralisation, they must be completed by complementary approaches, including quantitative and qualitative indicators to determine the real magnitude of decentralisation, to correctly grasp the trends at play and to accurately assess the impact and the outcomes of decentralisation.

Subnational fiscal autonomy

The concept of fiscal autonomy is not easy to assess. Fiscal autonomy concerns both sides of the budget, expenditure and revenues. Spending power may be limited, either because subnational governments act as “paying agent”, carrying out a centrally defined spending agenda with little or no decision-making power or room for manoeuvre, or because regulations, norms and standards or budgetary rules impose ceilings or compulsory expenditures. Revenue autonomy is also a complex issue and goes beyond tax autonomy. It depends, for example, on the extent of discretion in intergovernmental transfers (from earmarked and conditional transfers to general-purpose grants based on a formula), on the type of taxes (from shared taxes with no or little taxing power to own-source taxes on which subnational governments have a certain power to set rates and bases), on the ability to set the level of tariffs, user charges or fees or on the possibility of raising revenues from local assets.

Fiscal autonomy also includes the capacity to access external funding (e.g. by borrowing, engaging in public-private partnerships). It also includes the capacity to manage budgets, to hire and fire staff, to choose the modes for delivering local public services as well as control and reporting mechanisms, performance assessments, etc. Finally, equalisation instruments may also limit the fiscal autonomy of subnational governments, especially horizontal equalisation (from wealthy jurisdictions to the poorer ones, also called the “Robin Hood principle”).

Revenue autonomy is at the core of decentralisation. The proportion of grants versus tax revenues in subnational revenue and the power on these sources of revenue is key. A particular challenge is to set up the vertical distribution of tax revenues among levels of government and to determine which taxes to assign to subnational governments, under what criteria, and with which degree of discretionary power over tax bases and rates.

It is important to underline that there are no clear-cut frontiers between the different sources of revenue. Rather, it 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) (Figure 2.5). Even in this case, the ability to act on rates and bases may also be regulated and restricted, reducing subnational government taxation power (e.g. through the imposition of caps, exemptions imposed by the central government, etc.).

Between these two extremes, there are various gradations from revenue sharing (general grants), tax sharing, local surtax (or surcharges) on national taxes such as personal income tax, corporate income tax or value-added-tax to tax revenue generation with more or less room for manoeuvre in terms of rates and bases (caps on rate for example or limited exemptions).

Figure 2.5. The continuum of subnational tax autonomy

The need to go beyond fiscal indicators to measure decentralisation

Fiscal indicators such as subnational shares in spending, revenue, and investment tell an important story about where money is spent, but do not fully capture all aspects of spending and revenue power. In particular, fiscal indicators can be misleading because decentralised expenditure can just be delegated expenditure made on behalf of the central government. In many countries, subnational governments have little spending autonomy, especially when they are the “paying agent” for the central government, for example to pay the teachers’ salaries or distribute social and welfare benefits to households, with no or little choice of how expenses are allocated (e.g. in Denmark in the social sector). Often, subnational governments do not have full autonomy and decision-making authority in their fields of responsibility, functioning sometimes more as agencies funded and regulated by the central government rather than as independent policymakers. Nordic countries, for example, are in practice likely to be less decentralised than they might appear from fiscal indicators (e.g. Denmark, Finland, Sweden). In many countries, a significant share of public spending takes place at lower levels of government, but this information on its own paints an incomplete picture of what subnational governments can actually do, autonomously, to affect the lives of those living in their territory.

Fiscal matters are one dimension of multilevel governance and they need to be complemented by other indicators to give a full picture of decentralisation systems. Several measures exist that address other dimensions as well.1 The Regional Authority Index (RAI) and the Local Autonomy Index (LAI) are comprehensive attempts to measure the real degree of power of subnational governments – beyond fiscal indicators. They provide a picture of multi-level governance, which is closer to reality than what is seen when looking at fiscal indicators only.

The Regional Authority Index takes the region as the unit of analysis and covers 81 countries along 10 dimensions annually from 1950-2010 (Hooghe, Marks and Schakel, 2010[25]; Hooghe et al., 2016[10]). The RAI distinguishes between tiers of intermediate governance (see also Chapter 3).2 The ten dimensions of the RAI include notably fiscal autonomy, borrowing autonomy, but also lawmaking, executive control (see Annex A). The Local Autonomy Index (LAI) follows, where applicable, the methodology of the RAI. The Local Autonomy Index (LAI) was developed for 39 European countries and it reports changes between 1990 and 2014.

Box 2.5. The Regional Authority Index and the Local Autonomy Index

The Regional Authority Index (RAI) tracks regional authority on an annual basis from 1950 to 2010 in 81 countries. The sample consists of all European Union member states, all OECD member states, all Latin American countries, 10 countries in Europe beyond the EU and 11 in the Pacific and South-East Asia. The unit of analysis is the individual region/regional tier. The dataset encompasses subnational government levels with an average population of 150 000 or more. Regions with a special autonomous statute or asymmetrical arrangements are also coded separately.

Regional authority is measured along ten dimensions: institutional depth, policy scope, fiscal autonomy, borrowing autonomy, representation, lawmaking, executive control, fiscal control, borrowing control and constitutional reform.

Primary sources (constitutions, legislation) are triangulated with secondary literature and consultation of country experts to achieve reliable and valid estimates. A regional data set contains annual scores for regional governments or tiers and a country data set aggregates these scores to the country level.

The RAI has proven to have a solid convergent content validity and has been used as a regionalisation or a multi-dimensional decentralisation measurement.

For the LAI, some adaptations had to be made to capture the specific characteristics of local government. For example, it is not appropriate to speak about non-deconcentrated local government or the endowment of an independent legislature because these aspects are parts of local self-government by definition (see the European Charter of Local Self-Government). Furthermore, more dimensions have been taken into account and some revisions of variables have been made (Andreas Ladner, 2016[26]).

Sources: Hooghe, L. et al. (2016[10])Measuring Regional Authority: A Postfunctionalist Theory of Governance, Oxford University Press, Oxford; Regional Authority Index,

Ivanyna and Shah have also developed comprehensive measures of the degree of decision-making at the local level, i.e. the level of government closest to the people (2014[9]). The dataset developed by Ivanyna and Shah covers 182 countries, and it captures institutional dimensions of political, fiscal and administrative autonomy enjoyed by local governments. These dimensions are then aggregated to develop a “decentralisation index” and are then adjusted for heterogeneity to develop a “government closeness index”. The analysis conducted on the basis of the index shows that decentralised local governance as measured by the government closeness index is associated with higher human development, lower corruption and higher growth.

A set of subnational spending indicators was developed in 2010 by the OECD Network on Fiscal Relations across Levels of Government, based on a detailed assessment of institutional, regulatory and administrative control which central government exerts over various subnational government policy areas (Bach, Blöchliger and Wallau, 2009[27]). Five categories were established: policy, budget, input, output and monitoring and evaluation autonomy. They allow a better assessment of the differences between spending indicators, and the degree of authority of subnational governments in these selected policy areas.

Highly diversified decentralisation systems in the OECD and around the world

Subnational governments in the OECD

In 2018, there were, in OECD countries around 136 800 subnational governments, in 9 federal and quasi-federal countries and 26 unitary states. The number of subnational governments varies greatly across countries (Annex B). This number is not necessarily related to population size or density: France and the United States have approximately the same number of subnational governments. The same applies to the Czech Republic (or Switzerland) and Mexico, which have roughly the same number of subnational governments, although the Czech Republic is 11 times less populated than Mexico (and Switzerland 15 times).

In total, France and the US account for 54% of all subnational governments in the OECD, followed by Germany, Spain and Italy with the next largest number of subnational governments. Twenty OECD countries account for less than 4% of all OECD subnational governments (Figure 2.6).

The high number of subnational governments reflects the high number of municipalities. There are around 132 300 municipalities in the OECD, 4 000 intermediate governments and 519 regions and state governments. Nine countries in the OECD have only one level of subnational governments, the one of municipalities. These countries are mostly small in terms of population and/or areas. Taken together, they represent 2.1% of the OECD total area and 2.7% of its population. Nineteen countries have two levels (municipalities and regions/states) and seven countries have three levels (municipalities, regions/states and an intermediate level between the two, e.g. départements in France, provinces in Spain, Kreis in Germany, counties in the United States). Italy belonged to this group until the 2014-15 reform, which abolished the provinces as directly elected subnational governments. Three-layer countries represent one-third of the OECD total area and half of its population.

Some OECD countries are characterised by particularly high municipal fragmentation, which makes the provision of local services inefficient and raises issues of equity in access to services, including varying quality. Municipal sizes vary greatly from one country to another as well as within one country. On average, the municipal size amounted to around 9 700 inhabitants in the OECD in 2017-18 (Figure 2.7), ranging from around 1 700 inhabitants in the Czech Republic to almost 225 000 inhabitants in Korea.

Figure 2.6. Number of subnational governments by country in the OECD in 2017-18

Source: OECD elaboration based on OECD (2018[28]), Subnational Governments in OECD Countries: Key Data (brochure and database),

Figure 2.7. Average and median municipal size in the OECD and the EU in 2017-18

Source: OECD elaboration based on OECD (2018[28]), Subnational Governments in OECD Countries: Key Data (brochure and database),

Most countries with large municipalities have a structured sub-municipal level that allows them to maintain a certain level of proximity and local democracy despite large municipal governments. These are for example characterised by civil “parish-type” municipal administrative subdivisions under public law and may have their own delegated budget and elected representatives (council, mayor). These may even have their own staff but do not have full local autonomy, as they are creatures of the municipality3. Such networks of localities exist in a variety of countries including Greece, Ireland, Korea, the Netherlands, New Zealand, Portugal and Slovenia.

Not only are municipal average and median sizes low in several OECD countries, but a fair number countries may also have a high proportion of very small municipalities, either in terms of population, geographic area, or both. In the Czech Republic, France and the Slovak Republic, more than 85% of municipalities have fewer than 2 000 inhabitants (Figure 2.8).

Figure 2.8. Municipalities by population size class in the OECD

1. Size-classes are slightly different: less than 2 499 inhabitants, 2 500 to 4 999, 5 000 to 24 999, 25 000 or more.

2. Metropolitan municipalities are not included to avoid double counting.

Source: OECD elaboration based on OECD (2018[28]), Subnational Governments in OECD Countries: Key Data (brochure and database),

Diversity in subnational government spending responsibilities

To get a sense of decentralisation systems, it is useful to see what fiscal indicators say in terms of subnational expenditures in the different policy areas. OECD countries have different levels of decentralisation measured either by the degree of spending decentralisation or by the tax revenues perceived by subnational governments.

Subnational governments represent a large share of public spending on average in the OECD. In 2016, subnational government expenditure accounted for 16.2% of gross domestic product (GDP) and 40.4% of public spending in the OECD. The importance of subnational governments in the economy is particularly evident when considering their role as employers. The lion’s share of public sector workers is employed by the subnational level, with 63% of government personnel expenditure undertaken by subnational governments. Subnational governments also play a key role in public markets through their purchases of goods and services for intermediate consumption and commissioning of public works. In 2016, they accounted for 49.5% of public procurement in the OECD.

Subnational government spending responsibilities vary from one country to another, depending on whether the country is federal or unitary, its size and territorial organisation, the degree of decentralisation and the nature of responsibilities carried out in certain sectors. In fact, some spending areas, such as education, social services and health, generate a greater volume of expenses than others because they involve significant current expenditure (e.g. social benefits, teacher and social worker salaries, and hospital staff wages). When subnational governments are in charge of these responsibilities, this automatically results in a high level of expenditure. Yet, this does not necessarily mean a high level of decentralisation. In some cases, such expenditures are delegated from the central government. Subnational government expenditure can also be constrained by regulations, norms and standards, or budget balance targets.

Figure 2.9. Subnational governments are key policy actors across OECD, 2016
Subnational government expenditure as a percentage of GDP and public expenditure

Note: 2015 data for Mexico, New Zealand and Turkey. IMF data for Australia and Chile.

Source: OECD elaboration based on OECD (2018[28]), Subnational Governments in OECD Countries: Key Data (brochure and database),

In federal countries, subnational expenditures are, in most cases, higher than in unitary countries because they combine those of the state and local governments. Canada stands out. Some unitary countries have a high level of subnational spending as well, such as Denmark, Finland and Sweden. By contrast, in countries where local governments have limited competencies, the level of spending is also limited (e.g. Greece, Ireland, Luxembourg, New Zealand and Turkey).

The breakdown of subnational expenditure by economic function provides a measure of the subnational government role in several areas (Figure 2.10). Education represents the largest spending area, accounting for 25% of subnational government expenditure on average in the 32 OECD (4% of GDP). Health is the second highest budget item, accounting for 18% of subnational expenditure (2.9% of GDP). General public services (administration) and social protection sectors are the third largest subnational budget item, both representing 14% of subnational spending (2.3% of GDP). Just after social protection comes economic affairs sector (primarily transport, but also commercial and labour affairs, economic interventions, agriculture, energy, mining, manufacturing, construction, etc.). This spending area represents 13.6% of subnational spending in the OECD on average (2.2% of GDP).

Figure 2.10. Subnational government expenditure by area, 2016

Note: Estimates from the IMF for New Zealand and Turkey. No data for Canada, Chile and Mexico.

Source: OECD elaboration based on OECD (2018[28]), Subnational Governments in OECD Countries: Key Data (brochure and database),

It is also interesting to look at the spending responsibilities across levels of government (Figure 2.9). In some areas, subnational governments are the main actors, accounting for more than 60% of public spending (housing and community amenities which include distribution of potable water, public lighting, waste collection; environment; recreation, culture and religion). In other areas, spending responsibilities are more likely to be shared with the central government. This is the case for education, economic affairs and transport, public order, health and social protection, although these sectors account for the largest shares of subnational expenditure. These averages mask wide variations from one country to another (OECD, 2018[29]).

Figure 2.11. Subnational expenditure as a share of total public expenditure by economic function

Note: No data for Canada, Chile and Mexico. For the United States, data showed in the function “housing and community amenities” include the “environment protection” function data. OECD average is unweighted. The total of public spending is non-consolidated.

Source: OECD elaboration based on OECD (2018[29]), OECD Regions and Cities at a Glance 2018,

Diversity in subnational public investment

In most OECD countries, subnational governments play a key role in public investment. Subnational investment represented 56.9% of public investment in 2016. However, national situations are very diverse. This ratio tends to be higher in federal countries than in unitary countries, although the role of subnational governments in unitary countries such as France, Israel, Japan and Korea is particularly high.

Figure 2.12. Subnational governments are key public investors across OECD, 2016
Subnational government investment as a percentage public investment

Note: 2015 data for Mexico, New Zealand and Turkey. IMF data for Australia and Chile.

Source: OECD elaboration based on OECD (2018[28]), Subnational Governments in OECD Countries: Key Data (brochure and database),

A diversity of financing models

There is a large variety of financing systems for subnational governments. Subnational funding models depend on a mix of criteria, including whether the state is federal or unitary, the degree and the type of decentralisation spending as well as economic, historical and cultural factors. The structure of subnational government revenue also varies greatly across countries (Figure 2.13).

In 2016, taxes represent the main source of revenues for subnational governments in the OECD (45%) before grants and subsidies (37%). Revenue derived from local public service charges (tariffs and fees), property income (sale and operation of physical and financial assets) and social contributions represented 15%, 2% and 1% of subnational government revenue respectively.

The share of tax revenue in subnational revenue varies greatly from one country to another. It is a particularly significant share in some federal countries, where tax revenue frequently derives from own-source taxation and tax-sharing arrangements between the federal government and state governments, as well as local governments in some cases. In the US, there is no tax sharing system between the federal and the subnational governments. In two other federal countries, Austria and Mexico, tax revenue – regardless of whether from tax sharing or own-sources – contributed less than 10% of subnational revenue in 2016. In some unitary countries, tax revenue made up more than 52% of local revenue in 2016 (e.g. in France, Iceland, Latvia, New Zealand and Sweden). At the opposite end, taxes amounted to less than 15% of local revenue in Estonia, the Netherlands, the Slovak Republic and Turkey.

The share of tax revenue is not an indication of tax autonomy, which depends on many factors – such as the right to introduce or to abolish a tax, to set tax rates, to define the tax base, or to grant tax allowances or relief to individuals and firms.

Figure 2.13. The structure of subnational government revenue varies greatly across countries, 2016

Note: Tax revenues in this figure exclude revenues from social security contributions, which are included in the OECD definition of taxes. Please see Section A2 of the OECD Interpretative Guide for further information. OECD26 and OECD9 respectively refer to unitary countries and federal countries.

Source: OECD elaboration based on OECD (2018[28]), Subnational Governments in OECD Countries: Key Data (brochure and database),

Countries can be grouped into four families based on their degree of subnational spending and tax level characteristics, which cut across federal versus unitary distinctions (Table 2.1).

Table 2.1. Categories of fiscal decentralisation by subnational expenditure and tax revenue

Spending decentralisation

(as % of GDP and % of total public expenditure)

Tax revenue decentralisation

(as % of GDP and % of total public tax revenue)

Highly decentralised spending

Australia, Belgium, Canada, Denmark, Finland, Germany, Japan, Spain, Sweden, Switzerland, United States

High tax revenues

Canada, Denmark, Finland, Germany, Iceland, Japan, Spain, Sweden, Switzerland, United States

Medium decentralised spending

Austria, Colombia, Czech Republic, Estonia, France, Iceland, Italy, Korea, Latvia, Mexico, Netherlands, Norway, Poland, Slovenia, United Kingdom

Medium tax revenues

Australia, Belgium, Colombia, Czech Republic, France, Italy, Korea, Latvia, Norway, Poland, Slovenia

Low decentralised spending

Chile, Greece, Hungary, Ireland, Israel, Luxembourg, New Zealand, Portugal, Slovak Republic, Turkey

Low tax revenues

Austria, Chile, Estonia, Greece, Hungary, Ireland, Israel, Luxembourg, Mexico, New Zealand, Netherlands, Portugal, Slovak Republic, Turkey, United Kingdom

Sources: Allain-Dupré, D. (2018[30]), “Assigning responsibilities across levels of government: Trends, challenges and guidelines for policy-makers”,; OECD (2017[31]), Subnational Covernments in the OECD: Key Data (brochure and database), OECD, Paris.

Table 2.2. Types of countries by level of decentralisation when measured by fiscal indicators
Most decentralised (Type 1) to most centralised (Type 4)

Type 1

High decentralised spending and high tax revenues

Australia, Belgium, Canada, Denmark, Finland, Germany, Japan, Spain, Sweden, Switzerland, United States

Type 2

Medium decentralised spending and medium tax revenues

Czech Republic, France, Iceland, Italy, Latvia, Norway, Poland, Slovenia, Korea

Type 3

Medium decentralised spending and low tax revenues

Austria, Estonia, Mexico, Netherlands, United Kingdom

Type 4

Low decentralised spending and low tax revenues

Chile, Greece, Hungary, Ireland, Israel, Luxembourg, New Zealand, Portugal, Slovak Republic, Turkey

Sources: Allain-Dupré, D. (2018[30]), “Assigning responsibilities across levels of government: Trends, challenges and guidelines for policy-makers”,; OECD (2017[31]), Subnational Covernments in the OECD: Key Data (brochure and database), OECD, Paris.

Subnational governments around the world

At the global level, the OECD-UCLG World Observatory on Subnational Government Finance and Investment identified approximately 523 000 subnational governments in 100 countries around the world, representing almost 6 billion inhabitants, i.e. 82% of world population and around 88% of the world GDP. There is a significant amount of diversity in subnational structures, which can affect how countries design and implement decentralisation reforms (Figure 2.14).

Figure 2.14. Subnational governments (SNGs) at the global level

Source: OECD elaboration based on OECD-UCLG (2016[3]), Subnational Governments around the World: Structure and Finance,

At a global level, subnational expenditure amounted to 9% of GDP and 24% of public expenditure in 2013-14 (OECD-UCLG, 2016[3]). The level of subnational spending is uneven among world regions (Figure 2.15) and countries (Figure 2.16).

Figure 2.15. Subnational expenditure as a percentage of GDP and of public expenditure at the global level by main regional areas

Note: All averages are unweighted.

Source: OECD elaboration based on OECD-UCLG (2016[3]), Subnational Governments around the World: Structure and Finance,

Figure 2.16. Subnational expenditure as a percentage of GDP and of public expenditure at the global level by country

Source: OECD (2018) Subnational governments in OECD countries: key data (brochure and database).

Subnational governments are also key public investors at the global level, accounting for 40% of public investment on average worldwide (1.5% of GDP). This indicates that worldwide public investment is a shared responsibility across levels of government, making its governance particularly complex. The OECD Council Recommendation on Effective Public Investment across Levels of Government, adopted in 2014, acknowledged the importance of better governance for public investment. It is relevant not only to OECD countries but also globally.

Figure 2.17. Subnational governments account for a large share of public investment worldwide

Source: OECD elaboration based on OECD-UCLG (2016[3]), Subnational Governments around the World: Structure and Finance,

Assignment of responsibilities: Great variation at the regional level, less variation at the local level

The assignment of responsibilities depends on many factors, including a country’s institutional structure. Some common trends in the assignment of responsibilities can be identified (Table 2.3). Overall, there is greater variation across countries in the distribution of competencies at the regional level and less variation at the local level (Allain-Dupré, 2018[30]). In contrast to local governments, which are often general-purpose, it is not uncommon for intermediate tiers to be deconcentrated, possess dual executives, or have more limited autonomy, particularly in unitary countries. There are also single purpose or special purpose local jurisdictions in several countries (e.g. for education, health and sanitation boards and agencies, and public/private utilities.

Table 2.3. Breakdown of responsibilities across subnational government levels: A general scheme

Municipal level

Intermediary level

Regional level

A wide range of responsibilities:

● General clause of competence

● Eventually, additional allocations by the law

Community services:

● Education (nursery schools, pre-elementary and primary education)

● Urban planning and management

● Local utility networks (water, sewerage, waste, hygiene, etc.)

● Local roads and city public transport

● Social affairs (support for families and children, elderly, disabled, poverty, social benefits, etc.)

● Primary and preventative healthcare

● Recreation (sport) and culture

● Public order and safety (municipal police, fire brigades)

● Local economic development, tourism, trade fairs

● Environment (green areas)

● Social housing

● Administrative and permit services

Specialised and more limited responsibilities of supra-municipal interest

An important role of assistance towards small municipalities

May exercise responsibilities delegated by the regions and central government

Responsibilities determined by the functional level and the geographic area:

● Secondary education or specialised 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:

● 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 (in federal countries)

Source: OECD (2017[32]), Making Decentralisation Work in Chile: Towards Stronger Municipalities, OECD Multi-level Governance Studies,

In many countries, the municipal level tends to manage community services. Municipal responsibilities are not always precisely defined. Regulations often refer to the general clause of competency or “subsidiarity principle”. This principle gives local authorities the explicit freedom to act in the best interests at the local level. In this case, laws rarely limit or specify local responsibilities but rather enumerate broad functions, unless a particular responsibility is devolved by law to another government level.

In two-tier subnational government systems, the regional level usually provides services of regional interest because it operates at a larger scale. These services benefit from economies of scale, generate spillovers, involve redistribution and are required to meet the same standards across the jurisdiction. The regional tier may also facilitate co-operation and strategic planning.

In three-tier systems, the breakdown can be complex, sometimes resulting in duplication, overlap and co-ordination challenges. However, over recent decades, the intermediate level has lost many of its powers and responsibilities in favour 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.

It is necessary to make the distinction between competencies and functions. For each responsibility area, different key functions can be distinguished: regulating, operating, financing and reporting. Regarding the financing function, another distinction can be made between current expenditure and investment. In the OECD, health, education and social protection and law enforcement draw heavily on subnational expenditure, with subnational governments in charge of paying medical staff, teachers, social workers and police officers or providing social benefits on behalf of the central government. Often, subnational governments act as “paying agents”, carrying out these delegated functions with little or no decision-making power or room for manoeuvre, and these spending responsibilities can represent a great burden on their budget.

Rather than a clear-cut separation of responsibilities, most responsibilities are shared across levels of government and the trend toward shared responsibilities has increased over the past decades (see Chapter 3). Because subnational governments are embedded in national legislative frameworks, truly exclusive competencies rarely exist, even in federal countries. Shared competencies emerge either through explicit legislation or through residual policy acquisition. The need to share responsibilities may arise for functional reasons – as is common between municipal and regional tiers around issues of transport and infrastructure, environment and water, culture and tourism, communication, or economic development. It may also arise for financing reasons such as for social services. This mutual dependence requires a clear assignment of functions, a clear understanding of who does what by all parties, and well-developed co-ordination mechanisms (see Chapter 5).


[30] Allain-Dupré, D. (2018), “Assigning responsibilities across levels of government: Trends, challenges and guidelines for policy-makers”, OECD Working Papers on Fiscal Federalism, No. 24, OECD Publishing, Paris,

[26] Andreas Ladner, N. (2016), “Measuring Local Autonomy in 39 Countries (1990–2014)”, Regional & Federal Studies, Vol. 26(3).

[27] Bach, S., H. Blöchliger and D. Wallau (2009), “The Spending Power of Sub-Central Governments: A Pilot Study”, OECD Economics Department Working Papers, No. 705, OECD Publishing, Paris,

[6] Boex, J. (2011), Exploring the Measurement and Effectiveness of the Local Public Sector: Toward a Classification of Local Public Sector Finances and a Comparison of Devolved and Deconcentrated Finances.

[12] Chatry, I. and R. Vincent (forthcoming), A Global View on Subnational Governments in Asia: Structure and Finance.

[8] Department of Commerce (U.S. Census Bureau) (2012), The Economics and Statistics Administration of the U.S. Individual State Descriptions: 2012,

[7] Devas, N. and S. Delay (2006), “Local democracy and the challenges of decentralising the state: An international perspective”, Local Government Studies, Vol. 32/5, pp. 677-695,

[25] Hooghe, L., G. Marks and A. Schakel (2010), The Rise of Regional Authority. A Comparative Study of 42 Democracies, Routledge, London.

[10] Hooghe, L. et al. (2016), Measuring Regional Authority: A Postfunctionalist Theory of Governance, Oxford University Press, Oxford.

[9] Ivanyna, M. and A. Shah (2014), “How close is your government to its people? Worldwide indicators on localization and decentralization”, Economics,

[11] Oates, W. (1999), “An essay on fiscal federalism”, Journal of Economic Literature, Vol. 37/3, pp. 1120-1149,

[15] OECD (2018), Maintaining the Momentum of Decentralisation in Ukraine, OECD Multi-level Governance Studies, OECD Publishing, Paris,

[29] OECD (2018), OECD Regions and Cities at a Glance 2018, OECD Publishing, Paris,

[5] OECD (2018), Practical Methodological Guide for the World Observatory on Subnational Government Finance and Investment, OECD, Paris.

[28] OECD (2018), Subnational Governments in OECD Countries: Key Data (brochure and database),

[32] OECD (2017), Making Decentralisation Work in Chile: Towards Stronger Municipalities, OECD Multi-level Governance Studies, OECD Publishing, Paris,

[2] OECD (2017), Multi-level Governance Reforms: Overview of OECD Country Experiences, OECD Publishing, Paris,

[23] OECD (2017), OECD Territorial Reviews: Sweden 2017: Monitoring Progress in Multi-level Governance and Rural Policy, OECD Territorial Reviews, OECD Publishing, Paris,

[31] OECD (2017), Subnational Covernments in the OECD: Key Data (brochure and database), OECD, Paris.

[18] OECD (2016), OECD Territorial Reviews: Peru 2016, OECD Territorial Reviews, OECD Publishing, Paris,

[21] OECD (2015), The Metropolitan Century: Understanding Urbanisation and its Consequences, OECD Publishing, Paris,

[17] OECD (2014), OECD Territorial Reviews: Colombia 2014, OECD Territorial Reviews, OECD Publishing, Paris,

[14] OECD (2011), OECD Territorial Reviews: The Gauteng City-Region, South Africa 2011, OECD Territorial Reviews, OECD Publishing, Paris,

[13] OECD (2008), OECD Territorial Reviews: Cape Town, South Africa 2008, OECD Territorial Reviews, OECD Publishing, Paris,

[3] OECD-UCLG (2016), Subnational Governments around the World: Structure and Finance,

[19] Olowu, D. (2003), “Local institutional and political structures and processes: Recent experience in Africa”, Public Admin and Development, Vol. 23, pp. 41–52.

[24] Ostrom, E. (2010), Beyond Markets and States: Polycentric Governance of Complex Economic Systems, American Economic Association,

[1] Rondinelli, D., J. Nellis and G. Shabbir Cheema (1983), Decentralization in Developing Countries A Review of Recent Experience, World Bank.

[16] Shah, A. (2017), Horizontal Fiscal Equalization in Australia: Peering Inside the Black Box,

[22] Shah, A. and T. Thompson (2004), Implementing Decentralized Local Governance: A Treacherous Road with Potholes, Detours and Road Closures,

[20] Sharma, C. (2018), Emerging Dimensions of Decentralization Debate in the Age of Glocalization, Kurukshetra University, Kurukshetra, Haryana, INDIA,

[4] UNDP (1999), “Decentralization: A sampling of definitions”, Working paper prepared in connection with the Joint UNDP-Government of Germany.


← 1. See Ivanyna and Shah (2014[8]); Arzaghi and Henderson’s index of institutional decentralisation (2005); Brancati’s levels of political decentralisation (2008); Lijphart’s federalism index (1999); Treisman’s decision making decentralisation (2002); Woldendorp, Keman, and Budge’s autonomy index (2000).

← 2. Except in the case of special capital districts that fall in the regional level, the RAI does not code municipal governance. It also does not make assessments of what territorial units are doing with their authority, but focuses on formal multilevel governance arrangements.

← 3. Except in Portugal, where “freguesias“(Portuguese designation for “civil parishes”) are local authorities with the constitutional guarantee of self autonomy”.

End of the section – Back to iLibrary publication page