copy the linklink copied!4. Ways forward in the process of decentralisation in Portugal

This section starts by summarising some international experiences for a successful territorial and institutional restructuring. After this, the chapter presents three scenarios for enhanced decentralisation in Portugal. The scenarios presented in this chapter are by no means exhaustive, and should not be seen as pure alternative recommendations. In a sense, the alternatives are partly nested and complementary to each other. The options presented should be seen as basis for further discussion and possible further investigation.


This section focuses on the Portuguese model of decentralisation and on the alternative ways Portugal could promote decentralisation at the regional and local levels. The OECD view is that decentralisation and regional level governance in Portugal needs to be strengthened, in one way or another. Currently, the CCDRs are not sufficient to tackle the present and future problems faced by Portugal at the regional level. This chapter argues that there is no single way to solve this. Based on information on international practices and experiences from regional reforms implemented in selected peer countries, the chapter ends by presenting three options for ways forward for Portugal. The choice should be based on efficiency and democratic considerations.

copy the linklink copied!Making reform happen: Prerequisites for a successful territorial and institutional restructuring

While each country is a special case and the effectiveness of institutional reforms such as decentralisation vary considerably across countries, some preconditions for a successful regional reform can be identified. Since regionalisation reform is, in fact, a special type of decentralisation, the preconditions are however much the same as for other decentralisation reforms. Among these requirements, focusing on clear responsibilities and functions assigned to government levels, as well as strengthening subnational capacities and the overall co-ordination mechanisms, are fundamental. Other preconditions include allowing for territorially-specific policies and the possibility for asymmetric decentralisation, with differentiated sets of responsibilities given to different types of regions/cities, in particular the metropolitan areas. Such mechanisms are critical to providing institutional and fiscal arrangements that best respond to local needs. Some of the measures described below would be useful for Portugal irrespective of regionalisation reform.

Portugal could consider adopting experimenting and sequencing practices when implementing major multilevel governance reforms. The Polish experiences which were described in previous sections of this report provide a particularly interesting case in this respect. Also in Sweden (see Box 4.1), the government managed to reassign tasks from deconcentrated central government to regions with elected self-government and fiscal autonomy using experiments and voluntary approach. While this kind of reforming is slower than a “big bang” reform, it enables learning-by-doing and revising the decisions during the reform process, if needed. In Finland, the regionalisation reform, which eventually failed, was attempted with a very ambitious package (combining establishing a regional level, integrating health and social services and enlarging the role of private provision in basic healthcare). Judging by the Finnish experience, in any case, Portugal should probably avoid attempting a very radical regional reform in a very short time period.

Portugal could also consider intensifying and developing the data collection processes as well as setting up new databases on subnational government finances and service provision. In particular, the outputs and outcomes of the main public services could be reported and collected in more detail, but more detailed data on the input side would also be a welcome development. For example, there is currently no comprehensive information available on municipal spending by the Classification of the functions of government (COFOG) classification (municipality-specific data).

The benefits of the decentralisation and regionalisation reforms will be greater if they are well informed. The General-Direction for Local Government, the central government authority in charge of monitoring municipalities, would be a well-suited actor to play this role. Three main tasks could be specified:

  • The design and implementation of a full-scale business intelligence information system, reflecting a well-designed battery of key performance indicators. This would be relevant for municipal decision-makers and it would also permit a robust analysis of comparative efficiency among municipalities.

  • The enforcement of cost-benefit analysis for public investments at the municipal level.

  • Ensure the means, including human capital resources, to enforce an across-the-board implementation of systems of costs or analytical accounting in all municipalities.

High-quality information on Portugal’s subnational governments would enable enhanced steering and monitoring of subnational governments, both by central government and citizens/taxpayers. Moreover, with enough good quality data, the steering could focus on monitoring service outcomes instead of “micromanaging” input use and service production processes. Although new data collection processes and setting up additional databases would require extra resources, savings gained over time from reduced administrative work would be likely to exceed the costs of the initial investment.

It must be also emphasised that decentralisation should be seen as an entity and not a collection of separate policies. Therefore, reforms on particular areas of governance and decentralised system are likely to affect other policy areas as well. Establishing a regional government level is therefore bound to affect the central government, municipalities, intermunicipal units and parishes, etc. In the same vein, changes in spending assignments affect the revenue side and changes in one public service sector such as regional planning will affect other public services, and so on. In order to secure a successful regionalisation reform, all major components of multilevel governance should be on the table.

copy the linklink copied!
Box 4.1. Reforming the dual multilevel governance in Sweden

Sweden is one of the most decentralised countries in the world in terms of public service delivery and expenditure: about 25% of the country’s gross domestic product (GDP) is accounted for by subnational government expenditure and the subnational government enjoys extensive spending, taxing and decision-making autonomy. In Sweden, nearly all redistributive tasks have been devolved from the central government to counties and municipalities. Sweden has a two-tier system of subnational government, consisting of 20 county councils and 290 municipalities:

  • County councils (landsting) are run by directly elected assemblies and are mostly responsible for health services (80% of their budget). Counties may also engage in promoting culture, education and tourism. The responsibility for regional and local public transport is shared between the municipalities and the county councils. Ten county councils are responsible for regional development policy. In addition, the municipality of Gotland, which is an island, is with county council responsibilities.

  • Municipalities (kommuner) are also run by directly elected councils. Municipalities are responsible for basic and secondary education, kindergarten, elderly care, social services, communications, environmental protection, fire departments, public libraries, water and sewage, waste management, civil defence, public housing and physical infrastructure.

Until the late 1990s, County Administrative Boards (central government agencies) were responsible for regional development in each county. Since 1997/98, Sweden launched a reform to transfer these competencies from regional central government agencies (CABs) to self-governing counties. The national government has however not imposed a single model on the counties but, instead, different regionalisation options. It has promoted an asymmetric and bottom-up regionalisation as a gradual and experimental process (a laboratory of regionalisation).

The underlying idea is that decentralised policymaking leads to more innovation in governance. Therefore, from 1997 onwards, Sweden developed various regionalisation options in terms of political representation and responsibilities in different regions and in different phases: directly elected regional councils in the two “pilot regions” of Skåne and Västra Götaland, resulting from the mergers of two and three countries respectively; an indirectly elected regional council for Kalmar; and a municipality with regional functions for Gotland. The second wave (2002-07) started with the Parliamentary Act of 2002. This act made it possible for counties, if all local municipalities agreed, to form regional co-ordination bodies (indirectly elected bodies, i.e. in line with the Kalmar model) to co-ordinate regional development work.

The third phase of experimentation, since 2007, corresponds to a renewed bottom-up demand for regionalisation. It started with the publication of the recommendation for the future of the regional level, published by the Committee on Public Sector Responsibilities in February 2007. The committee argued for the extension of the “pilot region” model, which was assessed positively, the merger of current counties and the creation of six to nine enlarged regions in order address long-term challenges such as ageing. The reform has not yet been applied as such but this bottom-up demand for regionalisation persisted and, since 1 January 2015, 10 county councils out of 21 counties are responsible for regional development.

Sources: OECD (2012[1]), OECD Territorial Reviews: Skåne, Sweden 2012,; OECD (2017[2]), OECD Territorial Reviews: Sweden 2017: Monitoring Progress in Multi-level Governance and Rural Policy,

copy the linklink copied!Three scenarios for enhanced decentralisation in Portugal

Nearly 70 years of economic research and practical experiences accumulated from decentralisation reforms in different parts of the world have shown that there is no single way to decentralise successfully. In general, the results of decentralisation reforms depend on the objectives set for the reform, the overall policy environment and the way the reform is designed and implemented. Each country is a special case and, therefore, such reforms require tailor-made planning.

Furthermore, the benefits and costs of establishing a self-governing subnational government level are influenced by the tasks assigned, as well as the financing system and the governance model applied. The overall quality of public administration and central government long-term commitment are essential factors behind successful decentralisation reform.

Regional governments and regionalisation reforms form a specific dimension of decentralisation. Regionalisation can mean a policy that establishes a new government layer between central government and municipalities. But regionalisation can also mean policies that strengthen the already existing regional government, for example, with new spending and/or revenue assignments. The type of tasks that will be assigned to regional governments matter too, as regionalisation can mean decentralising from central government to regions or transferring tasks or revenue from local governments to the regional level, or both. Each of these approaches are likely to have different effects.

In the Portuguese context, full regionalisation means the launching of the administrative regions, which are already defined in the Portuguese constitution. While the proposals for the boundaries of the regions have varied over time, the latest discussion has focused on the five-region model. There seems to be less consensus on what tasks the administrative regions would be responsible for and how they would be financed. These are key issues that must be solved before practical preparations for administrative regions can start.

Since there is not yet a political commitment to adopt administrative regions, we have chosen to discuss three options for decentralisation and regionalisation in Portugal. While there are, of course, many potential alternatives available, these three approaches are the ones that are most often observed in other European countries, as discussed in previous sections of this report. These approaches can also be easily varied as more information on reform needs become available. The scenarios presented below are by no means exhaustive, and should not be seen as pure alternative recommendations. The three options are presented from the least to the most comprehensive. In a sense, they are partly nested and complementary to each other. The options presented here are also a basis for further discussion and possible further investigation.

Decentralisation and strengthened deconcentration without empowerment of regional level

If Portugal decides not to establish the administrative regions at this stage, there are still ways Portugal could strengthen the current decentralisation framework and the role of the regional level. In short, these measures comprise reorganising the allocation of tasks between the Commissions of Coordination and Regional Development (CCDRs) and other deconcentrated central government services, strengthening the capacity and fiscal autonomy of municipalities, clarifying the role of intermunicipal co-operation, intensifying the metropolitan governance in Lisbon and Porto areas, and rethinking the task assignments between different levels of multilevel government, notably municipalities, central government and the intermunicipal co-operative units (IMCs). Each of these potential measures is discussed briefly below.

Aligning the regional networks of deconcentrated central government departments with CCDRs

CCDRs are currently the main regional policy actors in Portugal. The main duty of CCDRs is to co-ordinate the deconcentrated central government in regions but their ability to deliver that task is currently rather limited. There are probably many reasons for this and the following raises the main points. First, the deconcentrated regional departments of the Portuguese line ministries are currently fragmented, forming a complex, intricate network of central government services. Reorganising the deconcentrated central government services so that they are more compatible with CCDRs would be the first step to improved regional policy co-ordination. For example, in the Ministry of Labour, there are regionalised social security services with 18 district offices, deployed in 441 local services for the public. The same ministry is responsible for the Institute of Employment and Professional Training (IEFP), which comprises 5 regional delegations (coinciding with NUTS 2) and 53 local employment centres that may have several municipalities as their area of influence or even infra-municipal areas. Similar situations may also prevail in other sectors, which calls for reorganisation.

Second, it should be ensured that the jurisdictions of the various consultative entities promoting vertical and horizontal co-ordination at the regional level do not overlap with the work carried out in CCDRs.

Third, the role of the presidents of CCDRs should be strengthened in relation to the presidents of sectoral deconcentrated government departments. This would improve the effectiveness of CCDRs.

Strengthening the capacity and fiscal autonomy of municipalities for enhanced service delivery

Despite efforts to decentralise public service delivery and financing, Portugal is still a much-centralised country. There are clear benefits that could be gained by stepping up decentralisation. As discussed above, Portugal would follow the development road taken by most developed OECD countries. As it appears that Portugal has already strengthened the legal and administrative status of municipalities and IMCs during recent years, now could also be the right time to go forward with fiscal decentralisation. It is therefore important that the decentralisation process, started late 2018, is continued and followed through.

It should be noted that a partial decentralisation, such as assigning only limited responsibility of the service delivery to municipalities, may not lead to the full benefits of decentralisation. Spending responsibilities should also be accompanied by local financing responsibility to secure the right incentives for subnational governments.

Enough resources for ex ante and ex post impact analyses should be allocated so that outcomes of decentralisation can be monitored and evaluated in a reliable way. Central government could also establish arenas and platforms for spreading information on best practices and continuous learning. Furthermore, capacity building programmes for municipalities and IMCs should be established to support and assist the local governments in achieving the desired level of services.

Clarifying the role of intermunicipal co-operation

Currently, the IMCs and MAs play only a marginal role in local public service delivery. This situation can be at least partly explained by the service menu of municipalities: as the tasks currently assigned to municipalities are not very demanding in terms of outcomes, there has been less interest for intermunicipal co-operation. Another explanation could be that a tradition for intermunicipal co-operation has not yet developed, at least not within the IMCs that have been formed. While this situation may change as more tasks are decentralised to municipalities during the coming years, the central government could nevertheless accelerate the process with financial incentives. One way to accomplish this could be to use the transfer system, for example by directing more transfers to IMCs, instead of municipalities, in case of public services with important externalities.

It should also be noted that if the administrative regions are not established, the need for utilising IMCs in regional policy will increase markedly. This aspect is discussed in more detail in the subsection “Decentralisation without full regionalisation by reinforced municipal and intermunicipal levels”.

Intensifying metropolitan governance in the Lisbon and Porto areas

It is worrying to note that growth in the Portuguese metropolitan areas, notably in the Lisbon area, has not picked up as in other regions of the country. In fact, as was discussed in previous sections of this report, productivity and employment have recently declined in Lisbon, making the capital city area one of two TL3 areas in Portugal to record a decline in these indicators. This development has probably many explanations and there are many potential solutions for this problem. Here, just one potential solution – namely enhanced metropolitan-wide policymaking – is discussed.

Currently, the Lisbon and Porto Metropolitan Areas (MAs) are organised as mandatory one-tier co-operative associations. The real decision-making authority in metropolitan areas is however held by the municipalities of these areas, together with central government. As a result, the current MAs lack both political and financial authority to truly influence metropolitan-wide challenges.

It seems clear that the ability of MAs to engage in real decision-making and financing of the metropolitan areas should be improved. The legal authority of MAs could be strengthened, but at the same time, it should be ensured that the metropolitan authorities have adequate capacity to take responsibility for the development in the region. Central government should give support to the MAs to build the capacity needed to take responsibility for the strategic development of the metropolitan areas.

As was discussed in previous sections of this report, it is not possible to present one universal metropolitan governance model for all purposes. The main alternatives, at least based on international experience, include:

  • one-tier fragmented government structures

  • one-tier consolidated government structures

  • a two-tier government model

  • voluntary co-operation/special purpose districts

  • creation of specialised metropolitan authorities (e.g. transportation, spatial planning, etc.).

The main question is, of course, linked to which organisation is actually responsible for developing the area. The current system could be strengthened by increasing the autonomy of the MAs (in relation to the municipalities in the region). One option could be to alter the MAs into subnational government units with elected councils and granting them taxation rights (for example property tax), as well as giving them important tasks of region-wide importance such as transport, environment, housing, land use and planning. In this setup, municipalities could concentrate on providing services that are purely local in nature. The management system based on specialised metropolitan authorities supposes the delegation of certain management functions of the municipalities and raises the problem of intersectoral co-ordination (e.g. how the transportation and housing policies are co-ordinated).1

Rethinking the task assignments between levels of multilevel government

Taking into account that Portugal already has several local government units at its disposal and considering also that, in practice, Portugal is still a relatively centralised country, there seems to be room for rethinking the service assignments between central government, municipalities, parishes, IMCs and MAs. This would require preparing a nationwide plan which would need considerable preparation time and planning resources. In order for such reorganisation to be successful and politically accepted, all main stakeholders (e.g. line ministries, representatives of municipalities, IMCs, central government agencies, research institutes) should be engaged in the process. This suggestion is of course linked with the proposal for stepping up decentralisation in Portugal presented above. On the other hand, such an in-depth study and rethinking the entity of assignments could just as well lead to the recentralisation of some tasks. Moreover, it is generally recommendable that countries evaluate their multilevel governance setting every five to ten years.

Decentralisation without full regionalisation by reinforced municipal and intermunicipal levels

The second policy strategy for Portugal could be to establish a compulsory regionwide intermunicipal association, which could take over the tasks of CCDRs. This co-operative regional body could also take responsibility for some tasks that are currently organised by IMCs. Some municipal tasks could also be included, but only if this can be justified by major externalities that could be internalised by such reorganisation.

The current IMCs could still be allowed to exist as an intermediate level. They could concentrate on services which have smaller service benefit areas than the one covered by the regionwide municipal association. The responsibility of metropolitan governance could be concentrated on special metropolitan authorities as described in the previous section. This is because regionwide authorities would be too big and heterogeneous to be effective in taking care of metropolitan needs.

The regional councils in this alternative would consist of members elected by municipalities in the region. The tasks of the co-operative regional governments could be limited to regional planning, managing EU funds and some other tasks with clear regionwide benefit areas such as environmental protection and regional roads. The co-operative regional governments should have decision-making autonomy on matters that belong to regional governments’ jurisdiction. The financing of the regions could be based on central government transfers and municipal member fees. Municipal financing share should be high enough to create incentives for municipalities to avoid the so-called common pool problem.

Some European Union countries, such as Finland and Slovenia, are currently implementing a form of co-operative regionalisation, which could be inspiring for Portugal. Both of these countries have in fact also planned reforms to establish a self-governing regional government level, but such proposals have not yet gained enough political support. Co-operative regionalisation can be seen as an alternative to full regionalisation but also as an intermediate stage towards regionalisation.

Decentralisation through complementary regionalisation and intermunicipal co-operation reforms

Establishing a full-scale regional government level as described in the Portuguese constitution, i.e. administrative regions, is the third and more comprehensive option. If the reform is well designed, establishing administrative regions could solve many issues that the other two alternatives presented in this section are not able to solve, or solve only partially. Compared with co-operative regionalisation, the benefits of a full regional government level include especially gains in local democracy, legitimacy, transparency and accountability. Administrative regions could also bring additional allocative efficiency, depending on the eventual tasks assigned to the regions.

It should be noted that establishing administrative regions would not exclude policies to strengthen decentralisation at the local level. Therefore, the proposals to empower municipalities and IMCs would still be valid as they would complement the measures to establish administrative regions.

The formal structure and governance model of the administrative regions, including the election of members to regional assemblies and formation of a regional board as well as other bodies, is defined by Law no. 56/91.

It should also be noted that establishing administrative regions does not exclude the possibility of having deconcentrated regional administration for purely central government services. In fact, many countries practice the so-called “dual model”, with elected regional governments and deconcentrated administrations. In many unitary countries such as in France (regional and departmental prefects), Poland (regional prefects) and Sweden (county governors as heads of county administrative boards), the central government representatives at the territorial level operate side by side with elected regional governments. However, it should be noted that, in general, the representatives of deconcentrated central government administration do not interfere or take part in the decision-making devolved to the regional governments. They do however ensure that subnational governments obey the law and regulations and, therefore, prefects and county governors must deal with complaints made by local citizens regarding subnational government services.

The tasks of the administrative regions could consist, at the first stage, of current CCDR tasks, with the exception of co-ordinating deconcentrated central government tasks (such as health or education). Nor would the administrative regions co-ordinate municipal tasks, except in cases where such tasks are closely related with regional tasks, or tasks which the regions have laid out in contracts with municipalities. Therefore, regional councils would have two main functions: regional development and territorial cohesion planning. The councils could also be responsible for EU Structural Fund programmes and their implementation.

The financing of the administrative regions could consist of central government transfers and own revenues. The administrative regions would need to be endowed with some budgetary and financial autonomy, as fiscal powers form an essential dimension of political and administrative capacity. The own revenues of administrative regions could, for example, be derived from one or several local taxes or a share of national tax revenues.

Portugal should avoid major overlaps with the tasks of administrative regions, IMCs and municipalities. While full prevention may be hard or impossible, the remaining overlaps should be well investigated so that potential problems can be anticipated and minimised. A clarification of assignments and jurisdictions would also be required between administrative regions and deconcentrated state administrations. Portugal may be inspired by practices in countries where a dual model (elected regional government and deconcentrated central government units) co-exist in regions. As was discussed above, such countries are, for example, France, Poland and Sweden.

As said, intermunicipal co-operation and administrative regions complement each other and hence should be promoted together as a driving force to improve local public services. Intermunicipal co-operation and regionalisation are two key mechanisms to both rationalise public expenditure and improve the quality of public services throughout the country.

The regionalisation alternative described here could be the first step in a series of public sector reforms. The next phase could be the reform of deconcentrated central government departments. Depending on political support for further decentralisation reforms, the administrative regions could be assigned tasks that are now arranged by central government in regions, such as example secondary education and specialised healthcare for example. Until then, current IMCs should co-exist with the administrative regions. They could concentrate on services which have a smaller service benefit area than the regions but larger than that of single municipalities. Once the administrative regions have been established and once there is agreement on further decentralisation, the role of IMCs could be diminished. It is likely though that voluntary intermunicipal co-operation, albeit less needed, would be still useful in certain cases.

The metropolitan regions should benefit from a special solution, as discussed in the previous sections. It should be noted, though, that the metropolitan governments should be enhanced in any case and should not depend on the solution found for the regional governance model. The administrative regions would probably be too big for metropolitan issues, as administrative regions need to look at the regions as a whole. It is likely that this would not leave them with enough interest and capacity to take into account the needs and problems of the metropolitan areas.

In the event regionalisation continues on the basis of administrative regions, the question of Mas’ relationship with administrative regions should be solved. Since administrative regions would not have supervisory power over the municipalities, it seems that the relationship between administrative regions and MAs should be upheld through dialogue and co-operation. If this solution is adopted, Nordic countries, notably Sweden, could provide an interesting example for Portugal.

The European comparison shows that many average-sized countries opt for a single level of intermediate government, in particular when municipalities or grouped municipalities are vast. From this point of view, Portugal could take inspiration from similar-sized countries, such as Sweden and Switzerland.


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

[3] OECD (2015), Governing the City, OECD Publishing, Paris,

[1] OECD (2012), OECD Territorial Reviews: Skåne, Sweden 2012, OECD Territorial Reviews, OECD Publishing, Paris,


← 1. See OECD (OECD, 2015[3])

Metadata, Legal and Rights

This document, as well as any data and map included herein, are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area. Extracts from publications may be subject to additional disclaimers, which are set out in the complete version of the publication, available at the link provided.

© OECD 2020

The use of this work, whether digital or print, is governed by the Terms and Conditions to be found at