# Chapter 3. Making multi-level governance work for more effective development

Despite several attempts at decentralisation reform since the 1990s, Thailand’s governance system remains highly centralised. Strong central government control over subnational governments has not led to uniform service levels or harmonised revenue bases. On the contrary, there are marked fiscal disparities between Thailand’s subnational governments. Thailand’s dual multi-level governance and high number of subnational governments (LAOs) results in a governance system that is complex and fragmented. There are several alternatives available for Thailand to tackle the current problems. A clear nationwide plan should be developed to prepare for reforming the subnational government structure, financing system, and spending and revenue assignments. LAOs should be empowered by enhancing their spending and revenue autonomy. Reorganising the current spending assignments between government levels should be another priority. Merger reforms or enhanced co-operation should be considered to build adequate capacity among subnational governments. A stronger own revenue base would contribute to self-rule and accountability among Thailand’s subnational governments. To that end, LAOs should rely on at least one important tax base: for instance, a local surtax on the personal income tax collected by the central government would be an option. In addition, Thailand should reform other sources of local financing, such as the property tax and the transfer system.

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.

Phase I of the OECD Multi-Dimensional Country Review of Thailand, published in April 2018, focused on the “five Ps” of the Sustainable Development Goals (SDGs): People, Prosperity, Partnerships, Planet and Peace. The report identified a number of areas for development in Thailand, in particular informality, inequality, productivity, the management of natural resources and reforming institutions. The report also touched on areas linked to decentralisation, identifying implementation of reform in this area as a key challenge for achieving better public service delivery in all parts of the country. The OECD recommended that Thailand empower local administrations in order to pursue decentralisation and improve the efficiency of public services.

Thailand’s 12th National Economic and Social Development Plan (2017-21) aims to further increase subnational fiscal autonomy (Box 3.1). According to the Plan, Thailand’s local administrative organisations (LAOs) will be given greater flexibility in terms of finance and human resource management.

The focus of Phase II of the Multi-Dimensional Country Review is in-depth analysis and recommendations for Thailand. This chapter is dedicated to multi-level governance and decentralisation. Its main aims are to provide an overview of Thailand’s multi-level governance from an international perspective, to describe the current challenges faced by multi-level governance and decentralisation in Thailand, and finally to provide some examples and suggestions for possible solutions.

Box 3.1. The 12th National Economic and Social Development Plan and the multi-level governance issue

The 12th National Economic and Social Development Plan (NESDP) highlights the inefficiency of the public management system as a significant obstacle to national development. As such, good governance has been embedded in NESDP as one of the six primary strategies of the 20-year National Strategy. The Plan provides guidelines for better decentralising power, enhancing public participation, and redistributing responsibilities among national, regional and local actors. The Plan proposes five main objectives.

Revision of the dual model of governance. Thailand has started several ambitious decentralisation reforms. The NESDP sets targets to enhance the efficiency of Local Administration Organizations (LAOs). This chapter suggests empowering LAOs further with the provision of specific local public services.

Clarification of roles, tasks and powers of each level of government. The Plan aims at redefining the areas of competence of authorities at the central, regional and local levels in order to minimise task redundancy. Responsibilities should be allocated at the level of government where provision is more cost-effective because of economies of scale and beneficial spillovers.

Step-up public sector’s performance. The Plan aims at placing Thailand’s public sector among the top performers in the ASEAN countries. This chapter proposes to strengthen the capacity of local authorities to cooperate and coordinate the provision of certain public services.

Reform of financing system. The Plan provides guidelines to improve the management of local revenues by both the central and local governments. This chapter proposes that Thailand should consider giving at least one important tax tool to LAO. Moreover, it encourages a redefinition of shared taxes and of the way general grants are redistributed.

Strengthening the subnational capacity for strategic planning and territorial development. According to the Plan, LAOs should formulate policies with some degree of autonomy. The Plan moreover encourages LAOs to team up with other stakeholders in accordance with good governance frameworks. This chapter recommends establishing formal and informal co-operation and co-ordination mechanisms between LAOs.

## Overview of multi-level governance in Thailand from an international perspective: A highly centralised country

This section provides a snapshot of multi-level governance and decentralisation trends in Thailand since the 1990s from an international perspective. It focuses on taxation and spending powers and the structure of LAOs’ spending and revenues.

### Thailand’s multi-level governance framework is highly complex

Decentralisation was mandated by Thailand’s Constitution in 1997 and implemented through decentralisation reforms in 1999, 2000, 2002, 2006 and 2008.1 As a result, Thailand’s current system of multi-level governance is based on a dual model: deconcentrated central government units and LAOs that enjoy certain levels of administrative autonomy as established by law (Figure 3.1).2

The deconcentrated central government units consist of 76 state Provinces (Changwat). The governors and other head officials of the state Provinces are appointed by the cabinet. The state Provinces are further divided into districts (Amphoe) and sub-districts (Tambon). The state provincial and sub-district officers monitor and supervise the LAOs in numerous ways. For example, provincial government officials have the right to approve the annual budget plans of LAOs, dissolve local councils and even dismiss local councillors.

The local administration sector is composed of several types of LAOs. There are currently 76 Provincial Administrative Organisations (PAOs), 2 441 municipalities and 5 332 Sub-district/Tambon Administrative Organisations (SAOs/TAOs).3 In total, there are currently 7 852 LAOs in Thailand. The LAOs have directly elected Councils and executives. The Bangkok Metropolitan Administration (BMA) and the City of Pattaya have a special status with greater powers and autonomy than the other LAOs.

While Thailand has made progress in implementing decentralisation, it remains a relatively centralised country. Despite efforts to decentralise power from central to subnational levels, Thailand’s public governance model can be characterised as a strongly centralised system. For example, the central government through its local officials can veto the decisions of LAOs, thereby weakening the self-governance of LAOs, municipalities and SAOs/TAOs.

### International perspectives on decentralisation in Thailand

Thailand is not the only country with a complex subnational governance model. Many other Asian countries have several government tiers with a large number of government units. Looked at from the aspect of government tiers, Thailand’s two-tier local administration4 model resembles the systems in Indonesia and Japan. Pakistan, the Philippines, the Republic of Korea and Viet Nam have three tiers; China has four tiers; and India has one tier in urban areas and three tiers in rural areas. In each of these countries, multi-level governance is more complicated and diverse than the number of tiers suggest. For example, countries may assign funding asymmetrically, favouring specific tiers of local government or strategic categories of spending. In some cases, the countries have multiple types of subnational government operating at the same levels with comparable legal status (Naoyoki and Morgan, 2017[3]).

The average population size of LAOs also varies considerably within Asian countries. In Thailand, the average population size of 8 000 inhabitants is one of the lowest among Asian countries, and very small compared, for example, with Indonesia’s 500 000 inhabitants (e.g. Table 3.1).

Table 3.1. Thailand compared with other Asian

Country

GDP per capita, current PPP USD (2014)

Total population, million (2014)

Urban population, %

Type of state

Number of subnational governments (2014)

Average size of local government, (inhabitants)

PRC

12 855

1 401.60

55.6

Unitary

2 852

491 445

Indonesia

10 517

254.5

53

Unitary

508

500 894

India

5 701

1 295.30

32.7

Federal

250 671

5 167

Japan

38 634

127.1

93

Unitary

1 741

73 004

Korea, Rep. of

25 998

50.4

82.2

Unitary

228

221 053

Malaysia

25 639

29.9

74.7

Federal

149

200 671

Philippines

6 969

101.8

44.4

Unitary

42 028

2 422

Thailand

14 242

67.7

50.4

Unitary

7 874

8 598

Viet Nam

5 629

93.4

33.6

Unitary

11 145

8 380

Note: Average size is calculated by dividing the total population by the number of subnational governments.

Source: (OECD & UCLG, 2017[4])

Thailand’s subnational government expenditure as a share of general government spending (17.7% in 2013 and 12.8% in 2016) and as share of GDP (3.9% in 2013 and 2.4% in 2016) is one of the lowest among Asia-Pacific countries and very low compared with OECD countries (Figure 3.2). For example, among Asia-Pacific countries the average shares for public spending and GDP are 37.5% and 11.5%, respectively (OECD & UCLG, 2017[4]; IMF, 2018[5]).

The autonomy of subnational government in Thailand seems weak from the perspective of revenues. The share of subnational tax revenue in total tax revenue is one of the lowest among Asia-Pacific countries (the share was 7.8% in 2013 and 8.4% in 2016) (Figure 3.3 – Panel A). Only Cambodia, the Philippines and New Zealand have lower values. Correspondingly, the share of grants in all subnational revenues is exceptionally high in Thailand (Figure 3.3 – Panel B) (OECD & UCLG, 2017[4]; IMF, 2018[5]).

Overall, Thailand’s subnational spending and revenue autonomy is modest at best. This conclusion seems to hold even when factoring in for the difficulties of comparing indicators of subnational spending and revenue shares across countries.5 The low spending shares and the fact that PAOs, municipalities and SAOs/TAOs are tightly regulated by central government officials, jointly lead to the conclusion that the degree of decentralisation in Thailand is comparatively low.

Due to strict central government regulation, subnational government debt in Thailand is low from an international perspective. In 2013, Thailand’s subnational government debt was 3.7% of GDP and 7.8% of general government debt (OECD & UCLG, 2017[4]). The share is lower in Indonesia and Korea, whereas in Australia, China and Japan the debt shares are much higher. In Japan, local government debt accounts for 37% of GDP and 15% of all public debt, while the same shares amount to 15% and 36% in Australia and 30% and 59% in China.

Since March 2018, LAOs have been allowed to borrow for three purposes: investment projects, debt restructuring and local government pawnshops. The debt service ratio must not exceed 10% of local government revenue (calculated by using a three year moving average of revenue data).6 LAOs must apply for central government permission to borrow. Permission is granted if they can provide a financial plan and follow a standard accounting system.

Thailand’s LAO investments constitute about 50% of general government investments and 1.5% of GDP (OECD & UCLG, 2017[4]; IMF, 2018[5]).7 These figures place Thailand in the middle group of Asia-Pacific countries.

## Thailand’s fiscal decentralisation framework: Wide-ranging tasks and underutilised own revenues

This section concentrates on fiscal decentralisation and subnational finance, elaborating on the expenditure allocation across levels of government as well as the local revenue structure. Some approaches and guidelines developed by the OECD are introduced in order to tailor them to the context of Thailand.

### Assignment of responsibilities across levels of government should be clarified

LAOs have been assigned a number of important tasks (Table 3.2). In principle, the tasks have been transferred according to the capacity of each level: the lowest levels of local administration are responsible mainly for local public goods and services, such as public health promotion, kindergartens and local infrastructure. Table 3.2 shows, however, that many assignments overlap among government tiers, which suggests that they are not as clear as they could be.

Thailand has been devolving numerous tasks to local government, but local autonomy remains weak. Assignment of responsibilities to different levels of governments is based on the decentralisation reforms of 2002 and 2008, which ambitiously mandated the devolution of 359 central government functions. However, the central government keeps tight control of local governments by defining the services that can be provided at the local level with numerous Acts. To date, 256 functions have been transferred to LAOs.

The responsibilities of the Bangkok Metropolitan Administration (BMA) are the same as those of the Provinces. The BMA is responsible, among others, for the implementation of local development plans and co-ordination between local authorities in its area, liquid and solid waste management, inter-municipal transportation, infrastructures, commerce and tourism, local culture, arts, and the protection and promotion of traditions.

Staff spending (30%), operating expenditure (23%) and investments (22%) constitute the major spending categories of LAOs, on average. The remaining categories are divided between subsidies (8%) and other expenditures (6%) (Figure 3.4). While these spending categories are the most important on average, it is worth mentioning that PAOs and the BMA deviate from this trend. In Bangkok, “other expenditures” constitutes the largest spending category (26%), whereas in PAOs investments (39%) are the most important spending type.

Table 3.2. Spending assignment across government levels

Central government

PAOs

Municipalities, SAOs, BMA and Pattaya City

Defence

Foreign Affairs

Justice

Police

Fire fighting

University

Higher education

Elementary and secondary education

Kindergarten

Public health

Public health and Curative services

Public health promotion

Pension payment

Elderly and child care centres

Infrastructure investment

Urban planning

Waterways and harbour maintenance

Water sewage maintenance

Maintain of local order, stability of communities and society

Planning and promotion of local commerce and tourism

Natural resources and environment management and protection

Art, culture and local wisdom

Source: (Laovakul, 2017[7]).

The main spending functions of LAOs comprise general administration, central budgeting, education, housing and local amenities (Table 3.3). The share of other expenditures such as health and social work is very small. While this classification of expenditure is not fully comparable to that used in international statistics on subnational government spending (COFOG), the high share of administrative expenditure and “central budgeting” in Thailand’s subnational spending raises questions.8 In OECD countries, on average, purely administrative expenditures usually constitute only a small fraction of subnational government expenditures.9 It is nevertheless clear that education and housing are by far the most important subnational government spending functions out of all public services assigned to LAOs.

The fact that the Provincial Administrative Organisations (PAOs) are often responsible for the same tasks as municipalities and TAOs/SAOs suggests an overlap in responsibilities. A significant overlap in assignments would be problematic, especially from the perspective of co-ordination, as uncoordinated assignments can lead to inefficiency in service delivery.

Overlapping assignments can also incentivise LAOs to avoid costs, for example, by transferring responsibilities to other levels of government. While to some degree this potential problem always exists in the case of shared service responsibilities, unclear responsibilities can intensify “passing the buck” behaviour. This issue is discussed in more detail in the next section.

In Thailand, the situation is further complicated by central government intervention in LAO decisions. Local offices of central government often operate in parallel with the local administration sector, which confuses service delivery.

Many of the functions assigned to Thailand’s LAOs currently seem to be underfunded or even unfunded. The main underlying reason is failure on the part of the current grant system to take into account the fiscal disparities and service needs of LAOs. In addition, as the revenue powers of Thai LAOs are restricted, and as the regional differences in tax bases are substantial, LAOs are often unable to draw on their own resources to compensate for the lack of central grants. Regional differences in tax bases reflect the differing production potential among Thailand’s regions, as discussed in Chapter 2.

Table 3.3. General administration and education are the main spending functions of Thailand’s LAOs
Spending of local administration by function, % of total expenditure

Function

Percentage

Education

19.16

22.65

Public health

4.17

Agriculture

0.63

Industry and public works

6.79

Commerce

0.75

Strengthening the community

1.13

Housing and community amenities

18.23

Social work

0.84

Maintaining domestic peace

1.75

Religion, culture and recreation

1.53

Central budgeting

22.37

Total local expenditure

100

Note: The data used to calculate the shares do not include the Bangkok Metropolitan Area.

Source: (Laovakul, 2017[7])

Underfunding and low subnational revenue capacities contribute to variations in spending and service provision between LAOs. An examination of spending differences between Thailand’s LAOs reveals marked differences in per capita spending between LAOs in all spending categories (see Figure 3.5 for Gini coefficients on per capita spending).10 This situation is a potential source of inequality in access to services and service quality.

### Transfers and shared tax revenues dominate subnational government revenues

Tax revenues comprise less than 10% of all revenues for LAOs. The main own tax revenue source is the building and land tax, which constitutes 16% of tax revenues (this calculation also includes tax sharing). Other local tax revenues, including local development tax, signboard tax, animal slaughter tax, bird nest collection tax, retail sale of cigarettes, and tobacco and gasoline taxes (only for PAOs), are considerably less important in this respect (Figure 3.6). In addition, PAOs and the BMA are allowed to collect hotel taxes. Other revenues of LAOs include fees, fines and permission charges, revenue from property, revenue from public utility provision and other revenues (Thailand Department of Local Administration).

Central government financing to LAOs consists of tax-sharing arrangements and transfers. The main source of shared tax revenue is value added tax (VAT) (28% of tax revenues), but central government shares also other revenues with the LAOs. The other shared revenues include business tax, alcohol and beer tax (23%), excise tax, motor vehicle tax and fee (14%), property registration fee, gambling fee, mineral fee, petroleum fee and other miscellaneous revenues (Figure 3.6).

The volatility of shared tax revenues is a potential source of unstable public service delivery. Value-added tax and other shared tax revenues are dependent on economic cycles, which poses a problem for ensuring stable public service delivery.11 This is also a potential source of pro-cyclical fiscal policy. In order to avoid sudden cuts in public service delivery, LAOs need to build financial buffers. However, these can be difficult to manage efficiently. The current system may also lead to subnational government budget deficits, which LAOs may find difficult to cover. While Thailand’s Public Debt Management Act states that central government will not be responsible for LAO debt, the volatility of revenues may nevertheless increase demands for central government bailouts.

Since the decentralisation reform of 1999, a certain level of revenue has been guaranteed from the central government to LAOs. In 2007, it was mandated that the central government would distribute no less than 25% of net revenues. This level has risen gradually with the current government target at 35%. As a result, the proportion of central government net revenue distributed to LAOs has increased from 13% in 2000 to 28% in 2017. The share is estimated to reach 29% for the fiscal year 2018 (OECD, 2018[6]).

A high share of central government financing of LAOs may lead to inefficient subnational government decision making (Kim and Blöchliger, 2015[8]). The major role of central government grants and tax-sharing arrangements in subnational financing may reduce the fiscal responsibility of LAOs. This may happen if LAOs expect central government intervention or bailouts in the event that problems occur. Such expectations can “soften” LAOs’ budget constraints and thereby distort their decisions on spending and borrowing. The central government transfers are currently divided in general-purpose, specific and conditional grants.

“General-purpose grants” could help close vertical fiscal gaps, but their definition leaves fiscal imbalances between LAOs unsolved. “General-purpose grants” are distributed proportionally to the difference between the expenditure of basic services and own revenue (that is, not including grants) of each LAO (Laovakul, 2017[7]; Laovakul, 2018[9]; Sudhipongpracha and Wongpredee, 2015[10]). The LAOs have had full discretion using the general grant. However, the formula used to define the general grant has not taken into account the revenue-raising capacities nor the service needs of the subnational governments. As a result, the general grant has been unable to reduce fiscal imbalances between LAOs. For fiscal year 2019-20 the general purpose grant formula will be changed to factor in the revenue-raising capacities of Provinces. To measure this capacity, Thailand’s Regional Development Policy Integration Committee will use indicators such as GDP per capita, household income, number of poor in each province, efficiency to implement the provincial plan and efficiency in local budget management.

“Conditional grants” are used to finance public services that appear high on national government priorities, notably in education or health care (e.g. school lunches, student achievement programmes, HIV-infected patients’ stipends, elderly people’s stipends). “Specific grants” are used to finance certain public services such as village water supply. The recipient LAOs must return the unspent specific grants at the end of the fiscal year to the Ministry of Finance.

In 2016, intergovernmental grants and shared revenues (mainly VAT but also other shared revenues) made up 90% of total local government revenues (Figure 3.7). Own revenues (locally levied revenues) therefore represented just 10% of the total revenues for LAOs.

There are major differences between Thai LAOs in terms of own revenues. In general, cities and towns are clearly better positioned to collect own revenues (Figure 3.8). Nevertheless, all types of Thailand’s municipalities – including the biggest cities – must rely on central government financing. It is also noteworthy that the revenues available to PAOs are smaller than those available to other types of LAOs. This is largely because PAOs lack a specific responsibility and instead oversee the whole area covered by their province. As a result, municipalities and even TAOs handle a much larger budget than PAOs. The capacity of PAOs to assist municipalities or TAOs when needed is therefore debatable.

The current financing system applied to Thailand’s LAOs fails to even out fiscal disparities and differences in the costs of public service provision arising from varying service needs in different parts of the country. The Gini coefficients of subnational revenues show that revenues collected from own sources are very unequally distributed. At the same time, funding from the central government (shared taxes and grants) is surprisingly equally divided (Figure 3.9). This observation applies for all types of LAOs. Central transfers and shared revenues do not equalise the differences in LAOs’ own revenue bases.

LAOs do not have strong incentives to tax their current tax bases. Localities with higher per capita income and therefore higher tax potential income receive a disproportionate amount of general-purpose grants (Sudhipongpracha and Wongpredee, 2017[11]; Sudhipongpracha and Wongpredee, 2015[12]).12 LAOs that could develop a relatively self-sufficient revenue base do not face the incentive to do so. The current system end up diverting resources from LAOs with weak fiscal capacity and that therefore would need more assistance.

The transfer system and shared revenue system should be changed in order to guarantee equal service provision in all parts of Thailand. The richest LAOs should be able to finance the bulk of their spending using their own revenue sources. Central transfers should target mainly areas with a low tax base and/or high service needs, such as a higher share of young or elderly people. As discussed in Chapter 4, pollution, poor environmental performance, and low access to water and sanitation may point at higher needs for water security.

This subsection discusses alternative ways to tackle the main challenges posed by decentralisation and multi-level governance in Thailand. The discussion focuses on issues raised in the previous sections of this chapter: the dual model of governance, spending assignments, economies of scale, the financing system and subnational capacities. Each of these is discussed in more detail in the following sections. As an introduction to this discussion, it should be mentioned that the OECD has identified 10 pre-conditions for making decentralisation work (Box 3.2). Among these, the focus on clarifying assigned responsibilities and functions, as well as strengthening subnational capacities and co-ordination mechanisms, is fundamental. Other pre-conditions are linked to allow some room for territorially specific policies and the possibility for asymmetric decentralisation, with differentiated sets of responsibilities given to different types of regions/cities. Such mechanisms are critical to provide institutional and fiscal arrangements that better respond to local needs.

Box 3.2. Ten guiding principles for effective assignment of responsibilities to make decentralisation work

1. Clarify the sector responsibilities assigned to different government levels: while inevitably most responsibilities are shared across levels of government, it is crucial to ensure adequate clarity and mutual understanding of the role of each level of government in the different policy areas to avoid duplication, waste and loss of accountability.

2. Clarify the functions assigned to different government levels: equally important as clarity in the assignment of policy areas, is clarity in the different functions that are assigned – financing, regulating, strategic planning, implementing or monitoring.

3. Ensure balance in the way different responsibilities and functions are decentralised: balanced decentralisation – (i.e. when the various policy functions are decentralised to a similar extent) – is conducive to growth. Ensuring balance in the way various policy functions are decentralised is essential to allow for complementarities across policies and integrated policy packages for effective territorial development approaches.

4. Align responsibilities and revenues and enhance subnational fiscal autonomy: the allocation of resources should be matched to the assignment of responsibilities assigned to subnational governments. Subnational governments should control a portion of subnational resources in order to promote their accountability.

5. Actively support subnational capacity building by the central government for the human, institutional and strategic dimensions of subnational governments: greater responsibilities at the subnational level need to be complemented with human resources capable of managing them. Too often this dimension is underestimated. Capacity development at the subnational level, particularly in poor or very small municipalities, must be actively supported with resources from the centre, and will require long-term commitment.

6. Build adequate co-ordination mechanisms across levels of government: since most responsibilities are shared, it is crucial to establish governance mechanisms to manage those joint responsibilities. Such tools for vertical co-ordination include, for example, platforms for dialogue, fiscal councils, contractual arrangements, conditionality, standing commissions and intergovernmental consultation boards.

7. Support cross-jurisdictional co-operation through specific organisational arrangements or financial incentives, to increase efficiency through economies of scale in investment or public service delivery.

8. Allow for asymmetric arrangements and pilot experiences: allow the possibility for asymmetric decentralisation, in which differentiated sets of responsibilities are given to different types of regions/cities, based on population size, urban/rural classification or fiscal capacity criteria. Ensure flexibility in implementation, allowing for pilot experiences in specific places/regions – and permanent adjustments through learning-by-doing.

9. Effective decentralisation requires complementary reforms in the governance of land-use, citizen participation and innovative public service delivery and governance.

10. Enhance data collection and strengthen performance monitoring: monitoring and data collection need to be carried out to monitor the effectiveness of subnational public service delivery and investments. Monitoring systems need to be designed in such a way to provide useful data for decision making and peer learning and with a limited number of indicators.

Source: (Allain-Dupré, Chatry and Moisio, 2018[13]; OECD, 2018[14])

### Revising the dual model of governance

In order to reap more benefits from decentralisation, Thailand could consider giving greater autonomy to LAOs. For instance, the interrelationship between deconcentrated regional central government authorities and LAOs could be altered gradually in ways that enhance the role of LAOs. At the same time, it is important to ensure that LAOs have adequate capacity and resources to take over both spending and financing responsibilities. In this respect, Thailand could benefit from the experiences of Poland, where a relatively successful decentralisation reform was implemented in the 1990s and thereafter (Box 3.3).

Thailand’s central government could create political and financial incentives for LAOs to deliver services in ways that support important national policy goals. Such a policy could replace the current “top-down” management of LAOs. Reforms on the subnational financing system and normative regulation on public service provision would be important components in this change. These are discussed in more detail in the following subsections.

Box 3.3. Decentralisation reform in Poland: A sequential approach to giving more autonomy to subnational governments

Poland provides a relatively successful example of the implementation of decentralisation reform. While it is not possible to measure precisely the effects of decentralisation on Polish society, Poland has performed well, for example, in terms of World Governance Indicators, compared with its neighbours and other former communist countries.

As a result of decentralisation reforms in the 1990s, Poland is now considered the most decentralised country in Central and Eastern Europe. Between 1995 and 2014, the share of subnational government expenditure in total public expenditure increased by more than 9 percentage points, from 23% to 32%.

As a result of the decentralisation reforms of 1990s and thereafter, Poland has transformed from a very centralised country to a decentralised one. One key factor behind the success of this reform was preparatory planning in local governance, which began during the 1980s under the Communist regime. The new model of local government was implemented relatively quickly after the political change (IBRD/World Bank, 2015). Another important factor in its success is the gradual and systematic way in which the reform was implemented. New responsibilities were transferred hand in hand with capacity building at the subnational level. In addition, new fiscal rules and territorial contracts were introduced to control and co-ordinate the decentralisation process. Considerable effort was taken to ensure that all stakeholders understood the goals of the reform and the likely outcomes. Training and information activities, often organised through non-governmental organisations, thus played a key role in implementation of the reform.

Although the decisions involved in the decentralisation reforms were taken very quickly after the collapse of communism, implementation took place in stages. Municipal self-governance was established first, and was followed thereafter by the introduction of regional authorities. Another important aspect of the Polish reform is that subnational government revenues were decided after the finalisation of spending assignments. Following the 2004 Act on Local Government Revenue, which modified the financing of subnational governments, subnational governments gained more financial autonomy, with a decrease in the share of central transfers. The use of earmarked grants was reduced, while tax sharing on personal income tax (PIT) and corporate tax revenues was introduced.

Source: (OECD, 2008[15]; Regulski and Drozda, 2015[16]; Kaufmann and Kraay, 2018[17]).

The overall complexity of Thailand’s decentralised system should be reduced. At present, both the municipalities and TAOs operate at the second tier of local administration. Municipalities are classified into three subgroups. The reforms on multi-level governance should aim to improve the accountability and transparency of subnational government decision making. Clarifying the subnational government structure and reducing the overlap between deconcentrated central government and the local administration are key measures in this respect.

Thailand should also consider the use of experimentation and staged implementation before applying reforms across the board. In some countries, such as Sweden (see Box 3.4), the government has managed to reassign tasks from the deconcentrated central government to LAOs using regional experiments and a voluntary approach. While this kind of reform is slower than a “big bang” approach, it encourages learning-by-doing and allows decisions to be revised during the reform process, if needed.

Box 3.4. Reforming the dual multi-level 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 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 have responsibility for regional development policy. In addition, the municipality of Gotland, which is an island, has the responsibilities of a County Council.

• 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, the County Administrative Boards (central government agencies) were responsible for regional development in each county. In 1997/98, Sweden launched a reform to transfer these competencies from regional central government agencies (CABs) to autonomous counties. However, the national government has not imposed a single model on the counties, but rather provided different regionalisation options (OECD, 2010). In essence, it has promoted an asymmetric and bottom-up approach to regionalisation as a gradual and experimental process – a laboratory of regionalisation.

The underlying idea is that decentralised policy making leads to more innovation in governance. From 1997 onwards, Sweden developed various regionalisation options in terms of political representation and responsibilities in different regions and in different phases. These included 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 phase (2002-07) was initiated by the Parliamentary Act of 2002, which made it possible for counties – if all local municipalities agreed – to form regional co-ordination bodies (indirectly elected bodies in line with the Kalmar model) to co-ordinate regional development work.

The third phase of experimentation, which began in 2007, corresponds to a renewed bottom-up demand for regionalisation. It started with the publication of recommendations for the future of governance at 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 had been assessed positively, the merger of current counties, and the creation of six to nine enlarged regions in order to address long-term challenges such as ageing. To date, the reform has not been applied in full, but the bottom-up demand for regionalisation has persisted. As of 1 January 2015, ten councils out of 21 counties are responsible for regional development.

Source: (OECD, 2018[14]; OECD, 2017[18])

Thailand should also consider strengthening the ability of the central government to co-ordinate with LAOs in ways that do not involve direct interference in LAOs’ operations. An important measure in this regard is the expansion of the statistical database on LAOs to better inform national decision-making. This will require building a comprehensive data collection on the main financing indicators and service costs in all LAOs. For instance, Chapter 4 discusses the data required to improve the management of water security.. In addition, there is a need to gather data on service needs and other major factors affecting costs, as well as main service outcomes.13 These data should be made freely available for all stakeholders and private citizens. The Thai government could benefit from examining the recent developments in Colombia in this respect (Box 3.5).

Box 3.5. Examples of subnational government databases

The Norwegian KOSTRA system is widely recognised in the OECD area as a best practice. KOSTRA is an electronic reporting system for municipalities and counties. It can publish input and output indicators on local public services and finances, and provide online publication of municipal priorities, productivity and needs. KOSTRA integrates information from local government accounts, service statistics and population statistics. It includes indicators of production, service coverage, needs, quality and efficiency. The information is easily accessible via the Internet and facilitates detailed comparison of the performance of local governments. The information is frequently used by local governments themselves as well as by the media and researchers. Although individual local governments could use KOSTRA more efficiently (e.g. by systematic benchmarking), the system has helped to facilitate comparisons of municipalities, thereby promoting “bench-learning” or “bench-marketing”.

Largely inspired by Norway’s KOSTRA model, Colombia has recently made considerable efforts to facilitate reforms on subnational governments by creating online databases on regional statistics. The National Planning Department of Colombia has prepared several new indicator systems to monitor territorial-level developments and the administrative and managerial performance of subnational governments. For example, the National Planning Department has jointly developed an online tool called TerriData with its domestic and international partners, in order to review and compare key data from Colombia’s departments and municipalities. TerriData currently consists of data on demographics, education, health, public services, public finances, and security for departments and municipalities. Investments in information bases contribute to the ability of central government to steer and monitor developments in subnational governments. In addition, the 2018 census – the first census to be undertaken since 2005 – will improve the quality of territorial statistics. Regular censuses are particularly important for territorial indicators. The resulting datasets are open to all to enable subnational governments and other stakeholders to benefit from this information.

Source: (Allain-Dupré, Chatry and Moisio, 2018[13]; OECD, 2017[19])

There are some examples of open data platforms already in use in Thailand which are designed for local project monitoring. For instance, one such database reports the progress of all the central government projects in each province (www.padme.moi.go.th). The data are provided by the Ministry of Interior’s provincial government officials and do not include LAO data. Another example is a database that reports the progress of all projects operated by LAOs in each province (e-plan.dla.go.th). The data are provided by LAO officials.

More detailed information on Thailand’s LAOs would enable a gradual policy shift from the current direct management of LAOs to a more indirect “steering by information” approach. Sufficient good quality data would enable the steering to focus on monitoring service outcomes instead of “micromanaging” input use and service production process. While data collection processes and setting up databases would require additional resources, the savings gained over time from reduced administrative work would likely exceed the costs of the initial investment.

### Reconsidering spending assignments

Reorganising current spending assignments between government levels should be a priority when reforming Thailand’s subnational government. Better clarified roles, tasks and powers for each level of government would contribute to the efficiency and transparency of public service delivery. In addition, reducing duplication and overlapping assignments between PAOs, municipalities and TAOs would enable more efficient central monitoring of LAO tasks.

Several countries currently grapple with similar challenges (Box 3.6). While the detailed recommendations to address assignment problems need to be tailored to each country’s situation, the alternative paths forward are often the same. The following paragraphs discuss the possible solutions.

Box 3.6. Examples of lack of clarity in assignments

In Brazil, there is a lack of clear division of responsibilities in a number of areas, such as health care, education, social security, welfare, agriculture and food distribution, environmental protection, sanitation and housing.

In Chile, at the municipal level, the current system of responsibilities is quite complex with few exclusive competencies and 13 shared competencies whose dividing lines between central or municipal interventions are not well established and are subject to change. Additional tasks are regularly attributed to municipalities without financial or technical compensation. Shared responsibilities, especially education and health, represent a significant burden on municipal budgets. Until the adoption of the law on municipal staff in 2016, municipalities lacked the option to adapt their human resources – either quantitatively or qualitatively – to these new functions.

In Colombia, the cumbersome allocation of responsibilities across different levels of government complicates the funding system making it difficult to assess whether a specific service, whose management corresponds to different layers of the administration, is appropriately financed.

In Finland, the 2007 reform on local government structures demonstrated that many provincial, regional and local actors share responsibilities, which engenders problem of co-ordination and possible overlapping.

In France, a 2017 report from the “Cour des Comptes” suggested a need to further clarify the allocation of competencies between departments and intercommunalités and to strengthen efforts to pool functions across municipalities. The report also regrets that the “compétence générale” clause was suppressed for regions and departments, but not municipalities, thus inhibiting clarification of the roles of the different tiers of government.

In Mexico, expenditure responsibilities are not clearly defined, with overlapping responsibilities in the two key areas of health and education.

In Peru, a lack of clear delineation of competencies has produced uncoordinated overlaps and a lack of ownership.

Source: (Allain-Dupré, Chatry and Moisio, 2018[13]; OECD, 2017[19])

The assignment of responsibilities within each public service area depends largely on the type of the service. While the economic theory on fiscal federalism cannot offer precise answers with regard to the choice of optimal fiscal assignment, some inferences can however be made (see Box 3.7 for a more detailed discussion). On the one hand, the lowest level of government are better placed to provide local public services, such as streetlights, local schools and child daycare. Local decision-makers have more information about local preferences, needs and costs of provision. On the other hand, higher levels of government may provide public services that entail economies of scale more efficiently and at a lower cost. Similarly, higher levels of government are often best suited to take care of public services with important redistributive features (e.g. specialised health care, secondary and higher education) and with considerable positive externalities or spillovers (e.g. major roads or main water pipelines).

Box 3.7. Assigning allocation, redistribution and stabilisation tasks

The traditional fiscal federalism literature divides public tasks into three branches: allocation, redistribution and stabilisation functions (Musgrave and Musgrave, 1980). The general consensus is that the allocation function – that is, the provision of public services – can be the responsibility of both the central and subnational level of governments. Responsibility is best applied at the central level when the services in question have no specific local interest. Subnational responsibility is justified when the benefits of the goods or services are spatially limited. Moreover, according to Oates’ “Decentralisation theorem” (1972), the subnational level is most suited to providing services and goods, unless the central government has a clear advantage in provision. An example of this would be considerable economies of scale in provision.

According to Musgrave and Musgrave (1980), the redistribution and stabilisation functions fall mostly under the responsibility of central government. In particular, central government is considered best suited to deal with monetary or fiscal policy. It is also widely accepted that the redistribution function should be carried out mostly at the central level, as the central government is more able to implement income redistribution from the rich to the poor and establish minimum standards of public services across regions (King, 1984[20]).

“Pure local goods” such as local infrastructure (streetlights, local roads), sewage, land use planning or basic education, are usually considered best suited for subnational government provision. It is nevertheless quite common for subnational governments to be involved, at least in some way, in the provision of services with redistributive features. In the Nordic countries, for example, even health, education and welfare services have been delegated from the centre to subnational governments. In the case of decentralised redistribution, the central government usually retains responsibility for co-ordination and ensuring the equity of citizens in different parts of the country. This can be achieved, for example, by using transfers from central to subnational governments, or through normative regulation (minimum standards), or both.

Ideal spending assignments between levels of government are almost impossible to accomplish in practice. This is mainly because there is always a certain degree of “path dependency” present in decentralisation and because radical reforms of multi-level governance are hard to bring about. In addition, the conditions in countries and regions often vary considerably due to historical, geographic, institutional, cultural and capacity-based reasons. However, although a common policy method is hard to come by, some trends are common to all countries in terms of the ways responsibilities are assigned across local and intermediary levels of government. Table 3.4 below provides an overview of features common to the assignment of responsibilities.

Since the capacity of subnational government units to accept the responsibility ultimately defines the spending assignment, asymmetric decentralisation can be a way forward. A phased in, or sequential, approach could then be worthwhile. Asymmetry in provision is more easily arranged in cases of “pure local public services”, because local preferences dominate and there is less need for central government interference. In cases of redistributive services, the choice between uniform and asymmetric provision is more difficult, because of aspects related to equity and the need to ensure a standard or minimum level of provision in all parts of the country, among others. The obvious alternative to asymmetric arrangement would be to strengthen the existing subnational governments.

Source: (Oates, 1972[21]; Musgrave and Musgrave, 1980[22]; King, 1984[20]).

In practice, decisions regarding the “best” division of fiscal responsibilities between central, intermediate and local governments varies from country to country and depends on the institutional setting, history and politics. The experience from practices adopted in various countries suggests, however, that these basic theoretical principles have in general been applied (Table 3.4).

Table 3.4. Assignment of responsibilities across subnational governments: 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, preelementary 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

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 treatment treatment

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.

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, 2016[23]).

While the complete elimination of overlapping assignments is not possible, several country examples demonstrate that successful multi-level governance reform on spending assignments can be accomplished (see Box 3.8 for a more detailed description).

Box 3.8. Examples of reforms on shared responsibilities

Denmark reformed its subnational government in 2007. The reform reassigned the tasks between levels of government, merged municipalities and reduced the number of intermediate governments (counties). As a result of the reform, counties were granted responsibility for the most demanding healthcare services including hospital services. Municipalities gained responsibilities for health promotion, social welfare and education. One of the aims of the structural reform was to reduce both the degree of shared assignments and the incentives for cost shifting between government levels. In order to tackle the latter issue, the municipalities were obliged to co-finance the rehabilitation services and training facilities provided by the counties.

Spain reformed its multi-level administrative structures in 2013 with the aim of clarifying municipal competencies and preventing duplication. The main goal of the reform was “one Administration, one competency”. The reform reduced competencies that were not attributed by law or delegated by other administrations without adequate resources (known as “improper competencies”). The reform intended to improve the definition of local competencies. It started by establishing a list of core competencies. Competencies not included in this list were referred to as “non-core competencies”. In any event, whenever competencies are delegated from upper levels of government to lower levels of government, it is compulsory to provide corresponding resources earmarked to finance the services involved.

In Japan, the 1988 Municipal Government Act provided a reference framework for the distribution of responsibilities across levels of government, establishing a distinction between mandatory responsibilities (including some that are shared with central government or delegated) and optional responsibilities. The 1999 Decentralisation Law eliminated opaque central decision making on local responsibilities and clarified competencies more generally. Subsequent waves of reform have continued to develop the goals of greater municipal autonomy, clear delineation of responsibilities and proper financing.

In the Netherlands, the Dutch decentralisation reform that was carried out during 2012-2014 aimed to reallocate competencies between the different levels of government, in particular by re-enforcing provincial and municipal responsibilities and by establishing simpler and clearer divisions of responsibilities between the different public actors, thereby helping to avoid the overlapping of functions.

Source: (Allain-Dupré, 2018[24]; The Ministry of Health, 2017[25]).

In addition to reducing overlapping assignments, the reorganisation of spending assignments should work to ensure that each government level is assigned tasks that they can deliver in practice. Accordingly, the reorganisation should take into account the current and short/medium-term financial and human capacities of the various LAOs. Building financial and human resource capacities, however, involves many aspects and may involve several measures, such as reforming the financing system, mergers, enhanced co-operation of LAOs and training the local public service administrators. These are discussed in more detail in the following sections.

### Enhancing economies of scale

Local fragmentation can make provision of local services inefficient. Thailand’s LAOs consist of nearly 8,000 units, which is high in international terms. There are large differences in per capita revenues and spending between LAOs. Hence, an important rationale for reforming Thailand’s LAOs is local fragmentation. In addition, the differences in fiscal and human resource capacities raise questions about equity in public service quality and access to services, as discussed in Chapter 2.

If full decentralisation is pursued, structural reforms could simplify administrative structure and improve spending efficiency. As previously discussed, Thailand’s municipalities and TAOs have both been assigned important tasks that require adequate capacities. Merger reforms or enhanced co-operation (voluntary or mandatory) can contribute to the capacity building of LAOs and therefore support measures to clarify assignments. However, efforts to centralise the current assignments of municipalities and TAOs or transfer them to PAOs would naturally reduce the need for radical merger reforms or co-operation.

Merger reforms may involve top-down decisions or bottom-up choices. Mandatory approaches have been chosen in Japan (during the “Great Sh wa”, 1953-1999), New Zealand (1989), Denmark (2007), Greece (2011) and Turkey (2008 and 2012). Forced amalgamations have been carried out either with strict pre-determined plans and numbers of subnational governments or with more flexible objectives. Targets have been used – and reached – in Japan and New Zealand. Municipal amalgamations in Finland, Iceland, Luxembourg, Netherlands, Norway and Japan (“Great Heisi” consolidation ongoing since 2006) followed a bottom-up approach (OECD, 2017[19]).

Voluntary merger reforms can be successful, especially when LAOs (often municipalities) see clear benefits from the boundary reform. That is why in some cases special grant schemes have been formed to create incentives to merge. In other cases, LAOs have been promised more important tasks after successful merger reforms. Municipalities may also conclude that their current or future spending assignments are too demanding for them without a merger.

Irrespective of the eventual choice between “top-down” or “bottom-up” approaches to mergers, the central government should take the lead in the planning process. Both voluntary and forced merger reforms can fail, especially when they involve much political bargaining that in turn might inflate costs and imply financing schemes. There should be a clear national plan on the optimal local administration structure (i.e. the number of municipalities and the average size of each type of LAO). The planning should be based on best available data on the fiscal capacities and factors that affect costs. In addition, the final decision between forced and voluntary merger should take into account best practices from other countries and the opinions of LAOs.

If a merger reform turns out not to be feasible, the alternative way forward is to increase co-operation between LAOs. While there are some risks involved related to transparency and accountability in intermunicipal or interregional co-operative arrangements, co-operation does provide a comparatively easy and flexible way to utilise economies of scale. Moreover, as discussed in Chapter 2, co-operation agreements do not rule out later merger decisions and are flexible enough for LAOs to withdraw when they do not benefit from them anymore.

### Reforming subnational government financing

If decentralisation reform is to proceed, the weight of own revenues in the financing of Thailand’s LAOs should be increased. At present, only around 10% of Thailand’s LAO spending is covered by own revenues. For LAOs to be truly autonomous and accountable, a considerably larger share of their financing should be based on own revenues. Thailand should therefore seriously consider implementing a comprehensive reform of subnational government tax bases.

Box 3.9. The subnational property tax in the OECD area

The property tax is a cornerstone of local taxation in many countries, but its implementation and management faces many obstacles. The merits of the property tax are regularly praised by economists: visibility, lack of tax export, productivity thanks to the stability of tax bases and a solid return on tax collection, lack of vertical tax competition by exclusive or priority allocation to the municipal level, implicit progressivity (property values rise alongside the revenue of their owners) and horizontal equity. These merits do not conceal the weaknesses and limits inherent in its practical application and management, which raise debates and encounter many difficulties. Such obstacles explain the fact that the importance of recurrent taxes on property in subnational tax revenue and GDP remains modest, although it varies considerably across countries.

In the OECD area, recurrent taxes on property represent 35% of subnational tax revenue on unweighted average, but between 90% and 100% of local tax revenue in Australia, Ireland, Israel, New Zealand and the United Kingdom, most of which are Anglo-Saxon countries (Figure 3.10). At the other end of the spectrum, recurrent property taxes represent a minor local tax revenue source (less than 10%) in the Nordic countries (Finland, Norway and Sweden), Estonia, Luxembourg, Switzerland and Turkey (OECD, 2016c). They account for 17-30% of local tax revenue in Hungary, Iceland, Japan, Korea and Poland. As a percentage of GDP, recurrent taxes on property range from 0.1% in Luxembourg to 3.1% in Canada and 3.3% in France, the unweighted OECD average amounting to 1.1%.

A property tax reform would be an important first step. The reform could focus on updating information on property values, which are now outdated (the values currently in use stem from the 1980s). Property tax is generally considered to be one of the least distorting taxes and can provide a stable revenue source for LAOs (Box 3.9). Thus, the use of property tax, which currently accounts for less than 1% of all tax revenues in Thailand, should be encouraged. It is therefore promising that a law on land and building tax is currently being examined by the National Legislative Assembly (NLA).

Going further, Thailand should consider allocating at least one important tax base, such as the personal income tax (PIT), to LAOs to strengthen their own revenue base. Property tax and other current local tax and revenue bases cannot be expected to fund a major share of subnational spending, especially because LAOs have been assigned significant tasks. Defining an optimal tax strategy involves taking several aspects into account, including economic efficiency, competitiveness, fairness, revenue performance and administrative cost. Since the main problem of Thailand’s subnational revenues is the narrow revenue base, the alternatives are approached here from the perspective of revenue performance. Taking into account the fact that LAOs already receive a share of VAT, this leaves one main alternative, namely the income tax base, or personal income tax (PIT).14

The most efficient way to collect local income tax in Thailand would likely be to define the local tax as a surtax on the central government PIT. In general, the yield of a local income tax depends on the coverage of the tax, the rate structure, and the effectiveness of tax assessment and collection. For efficiency reasons, tax collection and other administration should therefore remain with the central government. The LAOs should, however, be allowed to decide tax rates in their jurisdictions. In order to alleviate risk for vertical tax competition (which could lead to overly high income taxation), and to increase the co-ordination of income taxation, the central government could set upper and lower limits for tax rates. Such limits should also take into account the overall burden on tax payers and Thailand’s competitive position in the region.15 The use of PIT as a subnational revenue source is discussed in more detail in Box 3.10.

A local tax reform should be accompanied by adequate fiscal rules. The volatility caused by revenues from the current shared tax bases poses a problem for Thailand’s LAOs. A local PIT would also be prone to economic cycles, although probably to a lesser degree than VAT.16 Fiscal rules should therefore come with a particular focus on regulating local borrowing and budget balancing.

Box 3.10. Personal income tax in OECD countries: A significant source of revenue for subnational governments

In OECD countries, personal income tax (PIT) can represent a significant proportion of subnational tax revenue. In countries such as Denmark, Finland, Iceland and Sweden, where the share of PIT in subnational tax revenue ranges from 82% to 97%, personal income tax is a local own source tax, not a shared tax. In Denmark, the local PIT is collected by the central government together with the national PIT. In Finland, the base of the local PIT tax is determined by the central government, but municipalities have full control over the rate. In Sweden, subnational tax revenues come almost entirely from the local PIT, which is an own source tax, levied independently from the national PIT. Municipalities and counties have the same tax bases but decide independently to set their tax rate.

In Norway, the revenue from the PIT on ordinary income is collected by the municipalities for the central government, the counties and the municipalities. The split of PIT revenues between the three levels of government is determined by parliament as part of the national budget. The tax level is set annually by the Norwegian parliament as the maximum level of municipal income tax. In principle, counties and municipalities can lower the income tax rate for their municipality, but in practice all use the maximum rates. In Portugal, the two autonomous regions enjoy a certain degree of tax autonomy. They are able to retain nearly all of the PIT generated within their territories, and exercise strong control over the rate and base. Portuguese municipalities receive a local PIT surtax capped at 5% of tax receipts collected from local residents, although municipalities can decide to reduce this percentage. In Italy, the PIT is a shared tax and an own source tax. Part of the PIT receipts are shared and local governments can also choose to levy a surtax on the PIT.

In Japan, the government launched in 2002 an ambitious reform of all three major sources of revenues of subnational governments – local taxes, the local allocation tax and specific grants. This came to be known as the “Trinity Reform”. FY 2003 saw JPY 4 trillion worth of reforms to targeted subsidies. In the following year, FY 2004, JPY 3 trillion worth of revenues were transferred from the central government to subnational governments, with a portion of the income tax shifted to local governments as a personal residents’ tax. The local tax system in Japan is a complicated entity with many taxes and only limited autonomy for local governments. In 2012 OECD recommended Japan to start a comprehensive tax reform to reduce barriers to the effective use of existing powers to set local tax rates, and encourage local governments to rely primarily on existing local taxes on personal income, consumption and property, as they are relatively stable (NESDB/Thammasat University, 2009[1]).

Finally, in some unitary countries such as Latvia, Poland and Slovenia, the PIT is shared and accounts for more than 50% of subnational tax revenue. In Estonia, Lithuania, Romania and the Slovak Republic, the PIT was considered a shared tax between the central and subnational governments, until a recent reform of the System of National Accounts. Under the new methodology, PIT receipts have been reclassified as central government transfers and no longer as tax revenue.

Source: (OECD, 2018[27]; OECD, 2018[28]).

A transfer system would be important to account for the marked differences of income tax bases between LAOs. For instance in Thailand’s rural and remote areas, the share of agriculture is high, and taxable income in agriculture is often very low. Furthermore, in some LAOs (especially the big cities) the share of the unofficial workforce may be quite high, which can limit the efficiency of tax collection. At the same time, taxing powers could incentivize LAOs to improve their tax bases and tax collection, which in turn could contribute to solving the problem of unofficial labour force. Since it is likely that the per capita income tax revenues would vary considerably between LAO’s, the transfer system should be designed to take into account these differences.

Thailand should also consider reducing the weight of tax sharing in LAO revenues, as this clearly represents a source of volatility. As discussed above, the financing system reform should focus primarily on strengthening the LAOs’ own revenue bases. Some of the revenues currently allocated through revenue sharing could be used to finance the general transfer system.

There are strong arguments for reforming the transfer system as well. In particular, the weight of the general grant in the grant system should be increased relative to specific grants. The existing excessive use of specific and conditional grants in LAO financing is problematic for many reasons. For example, LAOs are currently incentivised to compete with each other for central government financing. This race does not encourage local communities to co-operate to resolve local and regional problems. In addition, the emphasis on specific grants incentivises LAOs to submit a large number of project proposals in order to secure local financing. This is an inefficient way to allocate central government funding to LAOs, because it increases administrative work, among other reasons.

In addition, specific and conditional grants may draw the attention of LAOs away from local needs and preferences. These grants may distort local decision making and eventually weaken allocative efficiency. For example, in Norway, the use of earmarked grants that aim to boost elderly care services was successful in this respect, but side effects included reduced spending on education, reduced childcare coverage and increased budget deficits (Borge, 2016). Earmarking may also weaken the transparency and accountability of local decision making compared to scenarios where LAOs are steered in accordance with legal obligations and are funded with general grants.

Thailand should also consider altering the general grant formula to better take into account the differences in tax-raising capacity and service needs among LAOs. Ideally, the grant system would incorporate models based upon the capacity to tax as well service costs. This would enable the system to support the poorest LAOs or those that have the highest costs, or both (see Box 3.11 for an example).

Thailand could benefit from the experiences of intergovernmental grant system reforms carried out in other countries. In Australia, a comprehensive grant reform was carried out in 2009, which replaced the old system based largely on specific grants with a formula-based general grant system. In Switzerland, the fiscal equalisation system was reformed in 2007. The new system consists of three elements: (i) horizontal tax revenue equalisation between cantons, (ii) federal aid to cantons with below-average tax-raising capacity, and (iii) a federal equalisation fund for cantons with high geographic-topographic or socio-demographic spending. To compensate for the additional federal spending linked to the vertical equalisation fund, the federal government reduced the cantons’ share in the federal income tax from 30% to 17% (Blöchliger and Vammalle, 2012[29]). In Sweden, the fiscal equalisation system has been based on general grants since 1990s and the Swedish government has put considerable effort into designing and implementing the transfer system. In Italy, a reform is ongoing to introduce grants that equalise the costs of provision of public services and call for more responsibilities of local governments (Box 3.11).

Box 3.11. Data-driven approaches to equalise fiscal capacity and spending needs across regions: the cases of Sweden and Italy

## Sweden: income equalising formula

The income equalisation grant equalises calculatory tax revenues between municipalities and between counties. The calculatory tax revenue is defined using the actual tax base per capita times the average tax rate. Local governments with a per capita tax revenue below 115% of the average tax receive a grant, and local governments with a tax above 115% of the average tax pay a fee according to a special formula. The formula can be written as follows (for receiving municipality/county):

Granti = tax ratej x (1.15 x tax basej – tax basei) x Ck

where Granti is the tax equalising grant for municipality/county i, tax ratej is the country average municipal/county tax rate, tax basej is the average municipal/county tax base, and tax basei is the tax base of municipality/county i. The result is multiplied with compensation rate Ck, which is 0.95 for municipalities and 0.9 for counties.

The main purpose of the income equalisation grants is to equalise differences in the local tax base. In 2015, there were large differences in the municipal tax bases, ranging from a minimum of SEK 191 500 per capita to a maximum of SEK 504 400 per capita. The majority of the municipal tax bases, however, are grouped near the mean tax base (SEK 239 934 per capita).

Municipalities/counties where the tax base is above 115% of the country average have to pay a contribution to the equalisation system. If the tax base is between 115% and 125% of the country average, the compensation is 0.60 times the exceeding amount of the tax base. For the part of the tax base that exceeds 125% of country average, the municipality pays 0.85 times the exceeding amount.

In 2015, a total of 38 out of the 290 municipalities had a tax base higher than 115% of the average tax base in the country, and had to pay a fee to the system. Twenty of these municipalities are located in the Stockholm County area.

## Italy: grants to equalise costs of provision of local public services

Italy has been introducing grants that equalise the spending capacity of local authorities. The aim is twofold. In terms of equity, equalisation grants guarantee that all authorities have enough resources to provide local services with uniform standards of quantity and quality. In terms of efficiency, equalisation grants cover only “standard costs” of provision of public services. The grants therefore stimulate higher accountability of local administrators, since expenditure levels above standard costs must be financed directly by local resources.

In Italy, equalisation grants are distributed according to a two-stage procedure. First, every fiscal year an equalisation fund is defined. A fixed amount of resources is then redistributed based on the estimation of “standard costs” for each local public service and municipality. The estimation of “standard costs” takes into account the type of services provided, the territorial features and the social-economic and demographic characteristics of the resident population. More in details, “standard costs” for each local administration i and service j are estimated using a five-step procedure.

First, the estimation of the following model provides the Standard Expenditure Needs for a service j (e.g., public roads maintenance):

where, is the local per capita expenditure of municipality i in service j, ${\mathrm{X}}_{ij}$ are variables that capture local preferences and needs (e.g. number of vehicles per resident) and environmental characteristics that impact the total productivity of input needed (e.g. morphological characteristics of the local administration). ${\mathrm{W}}_{ij}$ is a set of variables that capture the costs of input (e.g. the average staff expenditures per employee).

Second, once the coefficients of the above empirical model are estimated, the expected values ( ${\stackrel{^}{\mathrm{y}}}_{ij}$) of the current expenditure of each municipality are obtained as:

Third, the Theoretical Standard Expenditure Needs for local administration i and service j are computed according to the following formula:

where are estimated in the first step. ${W}_{j}^{*}$ is a set of targeted costs of input. The definition of these targets determine the reward for local administrations that are particularly efficient in managing resources.

Fourth, the allotment coefficient for each service j and local administration i is computed as:

${FS_CR}_{ij}=\frac{{N}_{ij}{\bullet FS}_{ij}}{{\sum }_{i}{N}_{ij}{\bullet FS}_{ij}}$

The total standard expenditure need of each local administration is determined by the aggregion of all the allotment coefficients. The final ratio gives the share of equalisation fund to be distributed to each municipality.

As of today, the distribution of equalisation grants based on standard costs applies to municipalities only. The evaluation of standard costs for regions is ongoing.

Source: (OECD, 2017[18]; Blöchliger and Vammalle, 2012[29]; Ballanti et al., 2014[30])

### Strengthening subnational capacities to finance infrastructure

The central government tightly control LAOs’ access to finance for local projects. As previously discussed, about 50% of all public investments in Thailand are carried out by LAOs. These investments are financed by LAO reserve funds, project grants and borrowing. However, most Thai LAOs rely on central government to finance their investments. Only the wealthiest LAOs have considerable reserve funds for investments. The use of these reserves is tightly regulated by the central government. In addition, the lenders for LAO borrowing consist mostly of institutions that are controlled by central government.

The grants for infrastructure investments are mostly earmarked and conditional. LAOs send their project proposals to the central government for project funding. Projects are ranked first by the line ministry and then by the Bureau of the Budget. If the project is approved for central government funding, the grant is usually paid as a conditional grant.

As discussed in the previous section, the current system may distort local decision making by drawing the attention of LAOs away from local needs and preferences. This can weaken allocative efficiency in two main ways. First, central government officials do not necessarily possess all the relevant information to select the most productive and effective investments. Second, LAOs may propose a project that is not necessarily the best option for the local community, but one that they think will be accepted by the central government.

The current financing system is also problematic from the perspectives of co-operation and co-ordination. As the financing system is based on applications sent by LAOs, this practice encourages LAOs to compete with each other for funding. The financing system may therefore discourage LAOs from co-operating with each other in investment projects. This limits the opportunities to utilise local innovation for public investment.

Thailand could ease the control over LAOs’ investments and giving more decision-making power to LAOs in order to better reap the benefits of decentralisation. Instead of maintaining direct control of LAOs, Thailand could address co-ordination issues by establishing new frameworks for horizontal and vertical dialogue and co-operation.

The OECD has developed a Recommendation on Effective Public Investment Across Levels of Government to help countries assess the strengths and weaknesses of their public investment capacity in a multi-level governance framework. The tool is designed to help countries apply a whole-of-government approach to setting priorities for improvement.

The OECD Recommendation consists of 12 Principles which have been divided into three pillars: (i) co-ordination of public investment across and within levels of government and policies, (ii) strengthening capacities for public investment and promoting policy learning at all levels of government, and (iii) ensuring proper framework conditions for public investment at all levels of government (Figure 3.11).

As autonomous entities, LAOs should be able to finance most investments with own reserves and borrowing. The central government could still provide grant funding for investments but to a lesser degree. For this to be possible, LAOs ought to have at least one strong tax base available to them, such as the PIT.

In addition to direct funding sources for investment (own savings, transfers, borrowing), Thailand’s government jointly with LAOs could develop local capabilities to utilise indirect funding sources (e.g. public-private partnerships, private equity, utilising joint public entities as responsible project organisations) (Figure 3.12). It should be noted, however, that some financing tools, such as public-private partnerships, require considerable expertise before they can be utilised. Such skills can be realistically expected only in the main metropolitan areas.

### Enhancing subnational capacity for strategic planning and territorial development

This section discusses briefly the subnational capacities for strategic planning and territorial development in Thailand, focusing on human resources management and strategic planning, as well as vertical and horizontal co-operation. Chapter 2 discusses forms of coordination between local administrations to integrate service provision in a metropolitan area. Chapter 4 shows how better management of water basins across jurisdictions can improve the allocation and quality of water.

The differing fiscal and human resource capacities available in varying sections of Thailand’s fragmented local administration sector constitute one of the main challenges of the multi-level governance system in Thailand. Nevertheless, lack of current human capacities at the local administration level ought not be used as an argument against decentralising. It is possible to develop human capacities with systematic training programmes (top-down), and with learning-by-doing and sharing information and best practices between LAOs themselves (bottom-up). Subnational government associations could have a role in this process (e.g. Box 3.12).

Box 3.12. Associations of subnational government can help disseminate information on best practices

Associations of subnational governments are often essential to public administration reform processes, as these intermediation bodies regroup information and provide stable negotiating partners for the government, thereby helping to reduce substantial information asymmetries and high transaction costs.

In Finland, the working groups in charge of drawing up the reform included members from the two coalition government parties and opposition parties, as well as members from the Finnish Association of Local and Regional Governments. This allowed the reform to gain wide political support, despite a change of government during the reform process. The organisation, representativeness and capacities of such bodies, and the quality of the relationships and trust between them and the government are key to the process; in contrast, their fragmentation may generate difficulties.

In France, the multiplicity of associations representing the interests of subnational governments at each level and within each level, while mirroring the diversity and vigour of local leadership, also makes it more difficult to build consensus and rally potentially diverging interests to uphold common positions.

Source: (OECD, 2017[19])

Thailand could also consider developing a strategy for LAO human resources management. The training of officials working in LAOs should be continued and intensified. In addition, statistical data on the public sector workforce should be developed further to help enhance human resource management.

Thailand could consider establishing formal and informal mechanisms for dialogue between LAOs and central government in order to advance the decentralisation process and to utilize all relevant information available. Many OECD countries have established co-operation and co-ordination mechanisms between central and subnational governments to empower subnational governments. The experience from these countries show that such arrangements can ease the planning and implementation of multilevel governance reforms. The instruments can be both formal and informal and they generally aim to promote dialogue and co-ordination across levels of government. The benefits of such structures include sharing experiences of best practices and preparing joint subnational proposals. The joint forums may also involve negotiations with the central level. In any case, a systematic dialogue between subnational governments and central government can lead to better understanding of the needs and problems at different levels of government and consultation in the design, implementation and monitoring of reforms (see Box 3.13 for some examples).

In 2017, Thailand introduced the Regional Development Policy Integration Committee as a new effort to integrate the policies and plans concerning regional development. The committee is chaired by the Prime Minister, comprises representatives from LAOs, private sectors, concerned government agencies, and NESDB as the secretariat. This committee could be a platform for LAOs to exchange ideas with the central government. The experiences of this Committee should be evaluated and, if found successful, the format could be extended to deal with other subnational government issues such as local tasks and funding.

Box 3.13. Co-ordination platforms for regional development and subnational investment

In order to ensure that their various levels of government take a more co-ordinated approach to regional development and public investment, many OECD countries make use of vertical and horizontal co-ordination platforms. These can include institutional mechanisms, co-financing arrangements, formalised consultation of subnational governments and platforms for regular intergovernmental dialogue. Practices in Australia, the Netherlands, New Zealand, Portugal and the United Kingdom provide relevant examples.

Infrastructure Australia was established in 2008 by Australia’s federal government to co-ordinate investments of national importance with Australian states and territories. Infrastructure Australia advises the national government on investment priorities in the transport, communication, energy and water sectors, and helps states identify infrastructure projects that align with national priorities. Infrastructure Australia assesses individual state or territory applications for funding under the Building Australia Fund, which is the country’s main mechanism for financing critical infrastructure projects.

In the Netherlands, the various levels of government establish their own vision documents – the Spatial Vision on Infrastructure and Spatial Planning at the national level, the Provincial Structural Vision at the provincial level and zoning plans at the municipal level. These documents serve as inputs to Area Agendas, which help all levels of government discuss and align their questions and projects in the physical domain (i.e. housing, industry, infrastructure, public transport, environment and water). Within the multi-year investment programme, each region has its own collective Area Agenda, which contains the co-ordinated vision, goals, questions and projects of the various regional government levels. Aligning the visions, goals and projects of each level of government leads to better solutions, greater efficiency and ultimately greater effectiveness. While formal discussions take place multiples times per year, decision making on the content of Area Agendas occurs at an annual meeting at the political level, with the outcome discussed in Parliament.

New Zealand’s Government Policy Statement establishes high-level priorities for transport investment, which are then implemented through the New Zealand Transport Agency (NZTA) in collaboration with subnational governments. NZTA officials work with each local authority to determine co-funding arrangements for the maintenance and renewal of the country’s regional and local roads (approximately 90% of all roads). Vertical co-ordination is largely confined to investment in Auckland. Auckland Council’s special plan sets out long-term priorities for public investment, and is designed to guide the investment decisions of central and local government, particularly in transport, as well as in social infrastructure (e.g. schools and hospitals).

Portugal’s Comissão de Coordenação e Desenvolvimiento Regional (CCDR) was created in 1979 for the purpose of planning. Currently, CCDR activities cover: spatial planning, promoting strategic and integrated regional development planning, monitoring the design and implementation of deconcentrated policies, and providing an opinion on the national government’s public investment expenditure programme (PIDDAC) at the regional level. In accordance with the EU Cohesion Policy, each region was requested to draft its own Regional Strategy 2020 under the direction of the CCDR, in order to improve collaboration among the CCDR, municipalities and the regional directorates of various ministries operating in the regions.

To improve local-level horizontal co-ordination, the UK government is encouraging the development of Combined Authorities, wherein cities and surrounding districts combine to create a representative entity with a legal status. The entity may share transport and economic development functions, as well as any other functions that their constituent authorities agree to share. To establish a combined authority, local authorities must develop a governance review that includes a recommendation for establishing such a legal structure for their area. The Greater Manchester Combined Authority, for example, was established in 2011. Other Combined Authorities were established in the Northeast, West Yorkshire, Sheffield and Liverpool in April 2014. The Authorities in Greater Manchester, Sheffield and West Yorkshire were offered additional powers via “devolution deals” in late 2014 and early 2015.

Source: (OECD, 2017[19])

## Policy recommendations

Goal to reach

Recommendations of the Multi-Dimensional Country Review of Thailand

1. Revising the dual model of multi-level governance

1.1. Enhance the autonomy of LAOs.

1.1.1. Move towards more indirect control and co-ordination of Local Administrative Organisations (LAOs) by the central government.

1.1.2. Strengthen the deconcentrated central government units’ capacity to consult and monitor LAOs. As these capacities accumulate, gradually diminish the direct interference of deconcentrated central government in subnational government decisions. In addition, strengthen subnational capacities to take over service tasks in order to gradually transfer all or most service tasks that satisfy local needs to LAOs.

1.2. Improve the accountability and transparency of subnational government decision making.

1.2.1. Create financial and political incentives for LAOs to meet the targets set by central government. This should be achieved by reforming the transfer system and developing enabling normative regulation that sets minimum standards, but leaves the responsibility of service provision to LAOs.

1.2.2. Local elections and elected councils and mayors of LAOs play important roles in supporting successful decentralisation reform. Local elections should be strengthened with core guidelines and arranged on a regular basis.

1.2.3. Strengthen the fiscal base of LAOs, especially the base for own source revenues.

1.2.4. Develop monitoring and prevention of corruption at all levels of government.

1.3. Strengthen the ability of central government to co-ordinate LAOs in ways that do not involve direct interference in LAOs’ operations.

1.3.1. Invest in an extended statistical database that covers all LAOs. Create a set of indicators that cover the financing, costs, availability and quality of subnational government services, for use in decision-making. Make all data and indicators publicly available.

1.3.2. Create a formal negotiation framework or body for subnational and central government representatives to discuss service level and quality targets, the financing of LAOs, and current and planned reforms. Establish informal forums for dialogue between central and LAOs in order to help promote co-operative arrangements and enable more co-ordinated actions.

2. Reconsidering spending assignments

2.1. Ensure an efficient and improved allocation of assignments between levels of government.

2.1.1. Clarify the assignment of responsibilities across government levels, taking into account the type of service (local, redistribution, externalities), especially in social services and education. In addition, use information on potential economies of scale and capacity differences to improve the efficiency of assignments. Make sure that there is no major duplication in assignments between government levels and that the resulting assignments are clear to all stakeholders.

3. Enhancing economies of scale

3.1. Ensure a stronger subnational government with visibly improved capacity to operate and take over devolved and delegated assignments.

3.1.1. Prepare a nationwide reform of subnational government structure. The reform should define the target number and size of LAOs at each government level. It should be based on decisions on spending assignments between levels of government and on the best available information on economies of scale, externalities, local democracy aspects and efficiency differences in service delivery. The reform should envisage both mergers and enhanced co-operation between LAOs. The reform should be developed and implemented in consultation with all main stakeholders and should be made public

3.1.2. A choice between forced and voluntary merger reforms should be made based on the national reform plan. If a voluntary merger approach is adopted, LAOs must have adequate incentives for entering mergers. Make sure that the resulting structure of LAOs supports the overall fiscal policy targets.

3.1.3. Promote and support co-operative arrangements between LAOs. If necessary, alter the legal system to allow for voluntary and obligatory co-operation.

4. Reforming subnational government financing

4.1. Strengthen own source revenues.

4.1.1. Carry out a reform of subnational government’s own source revenues. A key target of such a reform should be ensuring, at the macro level, that the majority of subnational government revenues are based on own tax and other own revenue sources.

4.1.2. Strengthen the property tax base by updating property values that are currently outdated.

4.1.3. Allow the subnational government level (at least the Provinces and municipalities) to tax income by means of a surtax on central government PIT. LAOs should be allowed to choose tax rates within limits set by the central government.

4.1.4. Simultaneously establish/update the fiscal rules on budget balancing and borrowing.

4.1.5. Reduce the weight of shared tax revenues and grants in the financing system.

4.2. Introduce a new grant system that supports equity between LAOs.

4.2.1. Increase the weight of the general grant in the grant system considerably at the expense of specific and conditional grants.

4.2.2. Reform the formula used to define the general grant to take into account differences in both fiscal bases and service costs.

5. Strengthening subnational capacities to finance infrastructure

5.1. Ensure that local information and innovation drive public investment

5.1.1. Strengthen subnational skills to invest using own revenue sources, borrowing and private sector financing. Make information available for different options for financing and set up training for subnational government administrators.

5.1.2. Ease central government control of subnational investment and instead establish new frameworks for co-operation and co-ordination of subnational and central government investments.

5.1.3. Strengthen abilities among major metropolitan governments to use indirect financing and management tools such as public-private partnerships, procurement, private financing and joint project management with several LAOs and private companies.

6. Enhancing subnational capacities for strategic planning and territorial development

6.1. Strengthen the human resource capacities of LAOs.

6.1.1. Revise and update strategies for human resource management. Collect statistical data on subnational human resources to support this approach. Continue and enhance efforts to train local administrators especially in strategic planning aspects.

## References

[24] Allain-Dupré, D. (2018), Assigning responsibilities across levels of government: Trends, challenges and guiding principles for policy-makers, OECD Publishing, Paris, https://doi.org/10.1787/22265848.

[13] Allain-Dupré, D., I. Chatry and A. Moisio (2018), “Asymmetric decentralisation: trends, challenges and policy implications (forthcoming)”, OECD Publishing, Paris.

[30] Ballanti, D. et al. (2014), “A Simple Four Quadrants Model to Monitor the Performance of Local Governments”, CESifo Working Paper No. 5062.

[29] Blöchliger, H. and C. Vammalle (2012), Reforming Fiscal Federalism and Local Government: Beyond the Zero-Sum Game, OECD Fiscal Federalism Studies, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264119970-en.

[5] IMF (2018), Government Finance Statistics (Database), https://data.imf.org/?sk=89418059-d5c0-4330-8c41-dbc2d8f90f46&sId=1435762628665 (accessed on 09 August 2018).

[17] Kaufmann, D. and A. Kraay (2018), Worldwide Governance Indicators (Database), http://info.worldbank.org/governance/WGI/#reports.

[20] King, D. (1984), Fiscal tiers: the economics of multi-level government., Allen &Unwin.

[8] Kim, J. and H. Blöchliger (eds.) (2015), Institutions of Intergovernmental Fiscal Relations: Challenges Ahead, OECD Fiscal Federalism Studies, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264246966-en.

[9] Laovakul, D. (2018), Country Profile: Thailand (unpublished memorandum).

[7] Laovakul, D. (2017), Fiscal Decentralisation in Thailand 2017 (PPT Presentation).

[22] Musgrave, R. and P. Musgrave (1980), Public finance in theory and practice, McGraw Hill Kogahusha.

[3] Naoyoki, Y. and P. Morgan (2017), Central and Local Government relations in Asia, Edward Elgar Publishing, https://doi.org/10.4337/9781786436870.

[1] NESDB/Thammasat University (2009), Decentralization and the Budget for Social Services at Tambon Administrative Level, NESDB/Thammasat University.

[21] Oates, W. (1972), Fiscal Federalism, New York, Harcourt Brace Jovanovich.

[14] OECD (2018), Making Decentralisation Work: a Handbook for Policy-Makers, OECD Publishing, Paris.

[6] OECD (2018), Multi-Dimensional Review of Thailand (Volume 1): Initial Assessment, OECD Development Pathways, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264293311-en.

[27] OECD (2018), OECD Tax Database - Explanatory Annex Part I: Taxation of Wage Income, OECD Publishing, Paris.

[28] OECD (2018), OECD Tax database (Database), http://www.oecd.org/tax/tax-policy/tax-database.htm.

[26] OECD (2017), Making Decentralisation Work in Chile: Towards Stronger Municipalities, OECD Multi-level Governance Studies, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264279049-en.

[19] OECD (2017), Multi-level Governance Reforms: Overview of OECD Country Experiences, OECD Multi-level Governance Studies, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264272866-en.

[18] OECD (2017), OECD Territorial Reviews: Sweden 2017: Monitoring Progress in Multi-level Governance and Rural Policy, OECD Territorial Reviews, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264268883-en.

[33] OECD (2016), Making the Most of Public Investment in Colombia: Working Effectively across Levels of Government, OECD Multi-level Governance Studies, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264265288-en.

[23] OECD (2016), OECD Regions at a Glance 2016, OECD Publishing, Paris, http://dx.doi.org/10.1787/reg_glance-2016-en.

[31] OECD (2014), Effective Public Investment Across Levels of Government: Principles for Action, OECD Publishing, Paris, http://www.oecd.org/effective-public-investment-toolkit/.

[32] OECD (2014), Recommendation of the Council on Effective Public Investment Across Levels of Government, http://www.oecd.org/regional-policy.

[34] OECD (2012), OECD Territorial Reviews: Skåne, Sweden 2012, OECD Territorial Reviews, OECD Publishing, Paris, https://doi.org/10.1787/19900759.

[15] OECD (2008), OECD Territorial Reviews: Poland 2008, OECD Territorial Reviews, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264049529-en.

[4] OECD & UCLG (2017), Subnational governments around the world: Structure and finance A first contribution to the Global Observatory on Local Finances. Country Profiles, https://www.uclg.org/sites/default/files/global_observatory_of_local_finance-part_iii.pdf.

[16] Regulski, J. and J. Drozda (2015), Building Local Government: Lessons of experience from the Polish transition, International Bank for Reconstruction and Development / The World Bank, http://pubdocs.worldbank.org/en/741171435014169639/Decentralization-EN.pdf.

[11] Sudhipongpracha, T. and A. Wongpredee (2017), “Fiscal Decentralization in Comparative Perspective: Analysis of the Intergovernmental Grant Systems in Indonesia and Thailand”, Journal of Comparative Policy Analysis: Research and Practice, Vol. 19/3, pp. 245-261, https://doi.org/10.1080/13876988.2016.1138659.

[35] Sudhipongpracha, T. and A. Wongpredee (2016), “Fiscal Decentralization in Comparative Perspective: Analysis of the Intergovernmental Grant Systems in Indonesia and Thailand”, Journal of Comparative Policy Analysis: Research and Practice, https://doi.org/10.1080/13876988.2016.1138659.

[10] Sudhipongpracha, T. and A. Wongpredee (2015), “Disequalizing equalization grant: An assessment of the relationship between equalization grant and local fiscal capacity in northeast Thailand”, Asian Affairs, https://doi.org/10.1080/00927678.2015.999515.

[12] Sudhipongpracha, T. and A. Wongpredee (2015), “Disequalizing equalization grant: An assessment of the relationship between equalization grant and local fiscal capacity in northeast Thailand”, Asian Affairs: An American Review, Vol. 42/1, pp. 1-21, https://doi.org/10.1080/00927678.2015.999515.

[25] The Ministry of Health (2017), Healthcare in Denmark: an Overview, http://www.sum.dk.

[2] World Bank (2012), Central-Local Government Relations in Thailand: Improving Service Delivery, World Bank, http://hdl.handle.net/10986/17362.

## Notes

← 1. The Constitution was reformed in 2007. In 2017, a new Constitution was issued following the coup in 2014.

← 2. Deconcentration should not be confused with decentralisation. Decentralisation implies delegating and devolving powers to autonomous subnational tiers. Deconcentration reorganises responsibilities within organisations. For example, some tasks may be shifted within central government from the main office to local offices.

← 3. Municipalities are divided into city municipalities, town municipalities and sub-district municipalities.

← 4. Subnational government tiers in Thailand – or Local Administrative Organisations (LAOs) - consist of provinces, municipalities and SAO/TAOs. Throughout the text, the terms “subnational government” and “LAO” are used interchangeably.

← 5. For instance, in some cases the spending share of subnational governments in total public spending can be quite high, but actual autonomy may be low. This may be due to central government control and delegated expenditure on behalf of central government, leaving little spending power to subnational governments.

← 6. The average revenue is calculated from the revenue projection from LAO’s Budget Document of the year that the borrow takes place, and the revenue from two years before.

← 7. According to IMF statistics and the UCLG/OECD publication, in 2012 local government investment in non-financial assets amounted to 1.5% of GDP. Central government investment in Thailand accounted for 3.0% of GDP at the same time. LAOs therefore represented 50% of general government investments in 2012. IMF figures for 2016 display slight variations with local governments representing 1.6% of GDP and central government 3.1% of GDP. Accordingly, LAOs are responsible for 52% of general government investments in this year.

← 8. Central budgeting is a special is a special spending category, which is reserved for emergencies.

← 9. In the OECD area, the two primary areas of subnational spending were education (25%) and health (18%) in 2016. These are followed by three spending areas that each amounted to around 14% of subnational spending: social protection, economic affairs and general public services (administration).

← 10. The Gini coefficient measures statistical dispersion. It is commonly used to measure inequality of incomes. A Gini coefficient of 0 would indicate perfect equality and a Gini coefficient of 1 would indicate full inequality among values. The Gini coefficient is based on the comparison of cumulative proportions of the population against cumulative proportions of income they receive.

← 11. A shared tax revenue system may also favour richer and densely populated localities because shared tax revenues are allocated mainly based on population size.

← 12. Tax potential is measured using per capita income.

← 13. Some examples of service need indicators in healthcare include demographics (age and sex), health status and socioeconomic status of population. Examples health care outcomes include premature deaths, patient injuries, infections and complications. In education, examples of need factors could be number of inhabitants in school age, adult literacy rate, number of pupils in need for special instruction. Education outcome indicators could be for instance OECD Pisa results, national (standardised) learning test results, number of students who finished primary/secondary education and share of dropouts.

← 14. VAT is comparable to local sales tax revenue tax, with the exception that the VAT rate is the same everywhere and that the central government takes care of collection.

← 15. Thailand and Viet Nam apply a 35% PIT, Indonesia is at 30%, Malaysia is at 28%, Laos is at 24% and Cambodia is at 20%.

← 16. Personal income tax base is usually less volatile than a VAT base. If subnational governments are allowed to set tax rates they can adjust they in accordance with economic cycles.

End of the section – Back to iLibrary publication page