Chapter 1. Challenges of subnational capacity development

Teresa Ter-Minassian

This paper, prepared for the 2nd KIPF-OECD-ADB Roundtable of the Network on Fiscal Relations in Asia (RoNFRA), draws in part on an earlier presentation by the author to the third meeting of the IDB’s Red de descentralización y gestión fiscal subnacional en América Latina y el Caribe.

Traditional (first-generation) theories of fiscal federalism (Tiebout, 1956[1]; Musgrave, 1959[2]; Oates, 1972[3]) emphasised the potential welfare and efficiency gains from fiscal decentralisation.1 Based on the key assumptions of benevolent governments, differences in preferences, and significant citizen mobility, they argued that expenditure functions should be assigned to the lowest level of government capable of internalising the benefits from those functions. This would improve preference matching and efficiency in the delivery of public services, because local governments were likely to know their citizens’ preferences better than the central government, and because citizens unsatisfied with their local governments’ performance could vote local officials out of office, or move to a different locality (“vote with their feet”).

The restrictive assumptions underlying such theories were subsequently challenged by second-generation theories (Qian and Weingast, 1997[4]; Oates, 2005[5]; Weingast, 2009[6])2 that emphasised political economy influences (including the risks of elite capture and governance failures) on the decentralisation process and its economic benefits.

More recently, both academics and policy makers have increasingly focused on the constraints that capacity limitations in subnational governments (SNGs) place on the effectiveness of decentralisation in ensuring a provision of public services that is more efficient and reflective of citizens’ preferences, as well as addressing equity concerns.3 This, in turn, raises several questions, including:

  • What are the main weaknesses in subnational capacity that hinder SNGs’ ability to efficiently and equitably deliver the goods and services for which they are responsible?

  • What are the incentives and resource limitations that affect SNGs’ willingness and ability to correct those weaknesses?

  • How can central governments (CGs) promote sustained capacity development efforts by their SNGs?

  • What role can horizontal intergovernmental co-operation play in improving subnational capacity?

This chapter explores such questions, drawing as much as possible on the (unfortunately limited) available empirical evidence from a range of countries at different levels of development. It also draws on the presentations made by representatives of a number of Asian countries at the Roundtable of the Network on Fiscal Relations in Asia (RoNFRA) for which this chapter was prepared, extending earlier work in OECD/KIPF (2019[7]).

The chapter is structured as follows. The next section defines subnational fiscal capacity and briefly discusses the limitations in available information on the state of such capacity. The third section explores the main weaknesses in subnational own revenue-raising capacity and in the design of intergovernmental transfers. The fourth section discusses weaknesses in SNGs’ capacity to carry out their assigned spending responsibilities, including flaws in planning, budgeting and executing spending programmes and public investments. The fifth section discusses the role of CGs in subnational capacity building. The penultimate section focuses on how intergovernmental co-operation can help address subnational capacity weaknesses, and the last section presents some brief conclusions.

As indicated above, increasing focus is now being placed in the literature and by policy makers worldwide on the fact that the effectiveness of decentralisation in promoting sustainable and inclusive growth depends on subnational fiscal capacity. The latter can be defined as subnational governments’ ability to:

  • secure own and transferred resources adequate to funding their assigned spending responsibilities in a reasonably efficient and fiscally sustainable manner

  • provide their citizens access to services and infrastructure of adequate quality

  • manage their budgets and financial and non-financial assets and liabilities responsibly and transparently.

Unfortunately, comprehensive, detailed, objective and internationally comparable information about subnational capacity needs is sorely lacking. The vast majority of systematic, indicators-based, international diagnostics about government capacities in revenue policy and administration (e.g. the International Monetary Fund’s Revenue Administration Fiscal Information Tool [RA-FIT] and the Tax Administration Diagnostic Assessment Tool [TADAT]4), and in expenditure and public financial management (e.g. Public Expenditure and Financial Accountability [PEFA]5 and Public Investment Management Assessment [PIMA]6), relate to central governments.

For the most part, qualitative country-specific sources of information are the sources available, including surveys of opinions of national officials7 and users of subnational services, and ad hoc studies covering one or more SNGs. Such studies can suffer from positive selection bias, as they focus more frequently on SNGs that provide better information (including performance indicators and/or cost data for spending programmes). These SNGs tend to be larger and richer jurisdictions in relatively more advanced countries and accordingly have comparatively more developed capacities.

Some countries (including the Philippines in Asia) have developed, or are in the process of developing, indicator-based systems of scoring financial and operational capacities of their SNGs, to identify major weaknesses, needs for support by the central government (CG), the appropriate pace of devolution of spending and revenue responsibilities, and subnational capacity to borrow responsibly.

Capacity needs vary not only across and among countries – reflecting, in particular, each country’s level of development and size, as well as the urban or rural nature of its jurisdictions – but also over time, reflecting growing challenges posed to SNGs by trends in the socio-economic environment, such as:

  • the dislocations caused by globalisation, and the related territorial shifts in demand and production, which affect both SNGs’ revenue bases and spending needs

  • migration flows, which can place increased burdens on social services (education, health and social assistance) provided by SNGs

  • changes in technology, which require upgrading the skills of subnational employees, and changes in education services delivered by SNGs

  • urbanisation and agglomeration trends, which lead to the emergence of mega metropolitan areas that include multiple jurisdictions, with related co-ordination challenges

  • reduced societal tolerance for large disparities in access to, and quality of, public services.

These challenges make subnational capacity development an ongoing imperative, requiring recurrent use of appropriate diagnostic tools, and the right incentives to design and implement appropriate corrective strategies.

It must be recognised that a number of political economy and governance-related factors can weaken subnational governments’ incentives to assess and improve their fiscal capacities, including:

  • factors that weaken the accountability of SNGs’ officials to their citizens, such as:

    • the capture of SNGs by local elites, which can hinder both their exploitation of own sources of revenues, such as property taxes, and their interest in providing quality services (e.g. primary health and education) to lower-income groups

    • limited involvement of the population in civic life

    • inadequate transparency of subnational budget operations

  • extensive discretion in intergovernmental fiscal relations, which softens the subnational budget constraint, and often results in favouritism toward SNGs politically aligned with the central government

  • a poorly functioning judiciary, which fosters impunity for tax evasion and corruption.

The effects of these and other governance–related weaknesses on subnational capacity are discussed in greater detail by Professor Paul Smoke in the next chapter of this volume.

Both first and second-generation theories of fiscal federalism recognise the benefits of a significant degree of autonomy for SNGs in deciding the level and composition of their own revenues.8 These benefits include:

  1. 1. the potential to increase overall national revenues by tapping sources (such as property taxes and user fees) that might be neglected or administered less effectively at the CG level

  2. 2. providing greater certainty to SNGs about their resource availability, thereby facilitating the preparation of more realistic budgets, and reducing volatility in the execution of spending programmes

  3. 3. promoting subnational fiscal responsibility, which tends to be undermined by SNGs’ reliance on gap-filling transfers or other bailouts by the CG (the so-called soft budget constraint) (Eyraud and Lusinyan, 2011[8]; Ter-Minassian, 2015[9])

  4. 4. facilitating the alignment of tax structure and design with local preferences

  5. 5. making more visible to electorates the cost of subnational spending, thereby increasing local officials’ political accountability and incentives to spend efficiently (provided that adequate transparency of their operations is ensured).

The literature also recognises, however, that there are significant economic, institutional and political obstacles to subnational own-revenue mobilisation. An important economic obstacle is the greater mobility of potential tax bases (goods and factors of production) within the national territory than across national borders, which increases the scope for tax evasion, and predatory tax competition (race to the bottom) among subnational jurisdictions. Another economic obstacle is the uneven distribution of the tax bases across the national territory, which implies that sole reliance on own revenues would result in excessive disparities in individual subnational governments’ ability to finance the provision of adequate levels of public services in key areas such as health, education and basic infrastructure.

There are also important institutional obstacles to subnational own-revenue mobilisation, including the better capacity of central than of subnational tax administrations to exploit economies of scale in the collection and enforcement of taxes; their generally better endowments of financial and human resources; and the fact that compliance costs for taxpayers operating in multiple subnational jurisdictions are magnified by the existence of differences in subnational tax legislation and tax administration procedures.

Finally, there are important political economy constraints to revenue decentralisation, in particular, central governments’ preference for maintaining control of the main tax bases, both to facilitate the conduct of revenue-based stabilisation policies, and to influence subnational spending decisions; and SNGs’ frequent preference for CG transfers (especially unconditional ones), to avoid the political cost of raising own revenues.

The balance between the benefits and costs of revenue decentralisation varies both across countries and over time. Given the host of factors affecting the degree of revenue decentralisation, it is difficult to find robust empirical explanations of it. It does not appear to be clearly correlated with the (federal or unitary) form of government; the level of development; the composition of gross domestic product (GDP); the degree of dependence on revenues from non-renewable natural resources; or even the degree of spending decentralisation.

The Asia region is characterised by relatively low levels of subnational tax autonomy. Although the share of tax in total subnational revenues (around 46%) is higher in other regions of the world, the majority of tax revenues are shared taxes on the level of which SNGs have no control, since the CG determines their bases, rate structures, collection and enforcement.

There is a vast literature discussing the criteria that should guide the assignment of specific forms of revenues to subnational (intermediate and local) governments [see, e.g. Ambrosanio and Bordignon (2006[10]); Bird (2010[11]); and Hagemann (2018[12])].

Theoretical considerations, as well as lessons from country experiences, suggest that desirable characteristics of such taxes include:

  • relatively low mobility of the tax base

  • the avoidance of distortions and risks of adverse spillovers on other jurisdictions, such as tax exporting, or predatory tax competition

  • a relatively even distribution of the tax base across the national territory

  • significant revenue-raising potential

  • low sensitivity to cyclical fluctuations and other exogenous shocks

  • the relative ease of administration

  • low compliance costs.

Although no potential subnational revenue handle meets all these criteria, some are generally recognised to be more suitable than others. At the intermediate (state or region) level of government, the most suitable revenue handles are surcharges on the national personal income taxes (PIT),9 retail sales taxes (RST) and excise taxes.

Surcharges on the PIT combine the advantages of taxing personal incomes (relatively low mobility of the tax base, and, if levied on a residence basis, conformity with benefit principle and low exportability) with reducing administration and compliance costs, and visibility (which are high in personal income taxation), a fact that increases their political acceptability. However, subnational PITs, or surcharges on the national PIT, are unlikely to produce substantial revenue in economies characterised by a high degree of labour informality, or by other weaknesses in the collection of the PIT at the national level. Therefore, they tend to be utilised mainly by advanced countries.

Compared to income taxes, subnational RSTs have the advantages of lesser cyclical sensitivity; more even distribution of the base; and possibly lesser visibility, which makes them more politically acceptable. However, they are difficult to enforce in countries (especially developing ones) where the retail sector is highly fragmented. They also tend to be regressive.

Turnover taxes (levied not only on sales to final consumers but also on inter-enterprise sales) tend to be favoured by SNGs for their high revenue potential, even at relatively low rates; relative ease of enforcement; and low visibility. Moreover, they tend to be more evenly distributed than is the case for income taxes and less cyclically sensitive. However, because of their cascading nature and exportability, these taxes entail high efficiency costs (Artana et al., 2012[13]).

Regional value-added taxes (VATs) do not suffer from the shortcomings of turnover taxes but have significant costs. These include: 1) limited subnational capacities to administer a multi-stage tax, especially if levied with multiple rates and multiple exemptions; 2) high compliance costs for taxpayers (businesses) operating in multiple states, if the regional VATs are levied on differently defined tax bases, or with different administration procedures; and most importantly, 3) the difficulties connected with the taxation of interstate trade. These problems are quite evident in the case of the state VAT in Brazil.10 Canada’s successful operation of dual (national and provincial) VATs reflects the advanced capacities of its national and provincial tax administrations and a well-established tradition of co-operation among them.11 India has only recently established a dual-type VAT (the goods and services tax [GST]); it is therefore too early to assess its performance.12

State-level taxes are also frequently levied on the consumption of specific (generally non-merit) goods and services, either on a stand-alone basis or, more frequently, as piggybacks on central government excises. Typical bases for subnational excises are gasoline, tobacco products, alcohol and soft drinks. Such taxes are increasingly also being levied, especially in metropolitan and/or touristic areas, on hotel occupancy and restaurant services. Their advantages are a reasonably good revenue-raising potential, low visibility and costs of administration (especially when collected at the point of production). They can also fulfil environmental or health policy goals.

In the traditional fiscal federalism literature, taxes on immovable (urban and rural) properties are generally presented as the ideal source of own revenues for local governments. This is the case because they are levied on immobile bases, and therefore are not exportable; they conform to the benefit principle, as they typically finance the provision of local services used by the property owners (or their tenants); can be mildly progressive (if assessed property values are kept reasonably close to current market values); and are generally13 less cyclically sensitive than income or consumption taxes. A limited number of local governments in OECD countries rely predominantly on property taxes (Figure 1.1).

However, economists like property taxes much more than taxpayers and politicians do (Bahl, Martinez-Vazquez and Youngman, 2010[14]). Their political unpopularity reflects several factors: the high visibility of the tax, which is typically paid in one or two annual instalments and may create liquidity difficulties for some taxpayers; the fact that the basis of the assessments is frequently not transparent, which opens scope for protracted judicial battles; and the fact that the difficulty of keeping property cadastres updated results in perceptions of horizontal inequities in the assessments.

Moreover, and importantly, property tax bases are typically very unevenly distributed across a national territory, with the bulk being concentrated in metropolitan areas; and property taxes are among the most challenging taxes to administer properly, especially in developing countries that are characterised by a high level of informality in real estate markets. Although technological advances have made the identification and registration of properties, and the updating of computerised property cadastres, easier than in the past, progress in these areas has been slow, especially in low-income countries, and in smaller municipalities. In addition, updating property assessments to reflect changing market values remains a challenge worldwide (Slack and Bird, 2014[15]). Moreover, the political economy factors mentioned above often discourage local officials from investing the resources necessary to keep property cadastres current.

A possible approach to addressing administrative challenges is assigning the responsibility for the administration of the cadastre to higher levels of government while maintaining local control over property tax rates and the enforcement of tax collections. This approach raises difficult issues of incentives (principal-agent relations between the levels of government involved), and empirical evidence on its effectiveness is not conclusive so far.

Various presentations at the Roundtable of the Network on Fiscal Relations in Asia confirmed that the above-mentioned weaknesses in the design and management of property taxes are common in Asian countries. Some, such as the Philippines, for example, are taking a range of steps to address such weaknesses, including tightening the enforcement of requirements for the updating of subnational property cadastres, and the approval of proposed subnational schedules of property market values by the national Finance Ministry.

Subnational governments also frequently levy taxes on movable properties, in particular automobiles. These are much easier to administer and are less controversial than real estate taxes. In setting rates, often there is a tension between environmental objectives (which argue for taxing less the more recent, less polluting models) and distributional ones (which support taxation based on the value of the car, implying lower taxes on older cars).

User charges are another suitable source of revenues for local governments. They can be levied to recover the cost of many services provided by these governments (e.g. water and sewerage, electricity, parking, garbage collection, urban public transport), and to contribute to the financing of other social services (such as education and health). They conform to the benefit principle and are largely non-exportable. Distributional concerns can be addressed by exempting – or levying the charge at reduced rates for – consumption levels below a minimum threshold. User charges can also help increase the accountability of local officials to their electorate for delivering public services of acceptable quality and minimising waste.

Many subnational governments, especially large cities, own assets (such as land, buildings and commercial enterprises) with significant revenue-earning potential (Detter and Fölster, 2017[16]), but relatively few (mostly in advanced countries) exploit such potential adequately. Most SNGs do not even have complete registers of such properties and reasonable estimates of their market values. Typically, realising their earning potential would require that their management be entrusted to independent, well-qualified agencies, and conducted on strictly commercial criteria.

It should be emphasised that even SNGs who have been assigned significant sources of own revenues may not have the right incentives to exploit them adequately, especially if they are subject to soft budget constraints, e.g. because of over-generous or highly discretionary transfers from the central government, or because of ineffective systems of borrowing controls.

They may then engage in predatory tax competition with other, especially neighbouring, jurisdictions through a proliferation of exemptions or other preferential treatments,14 or by keeping tax rates or user charges too low (race to the bottom). A possible remedial approach would be for the CG to set floors on the rate structure of taxes assigned to SNGs, and/or to limit the scope for the latter to grant an exemption and special benefits under these taxes.

Furthermore, SNGs are frequently hampered in realising their revenue-raising potential by a variety of limitations in administrative capacities. These limitations include, in addition to the above-mentioned weaknesses in the management of property cadastres:

  • cumbersome systems of tax collection, which raise taxpayers’ compliance costs

  • inadequate monitoring of, and slow reaction to, tax arrears

  • a high degree of discretion in assessments (e.g. of property values) and in the resolution of tax disputes, which widens the scope for favouritism and corrupt practices

  • low technical skills among tax officials

  • limited use of modern information technology (IT) systems.

Some of these weaknesses tend to be more prevalent in less developed regions and small, rural communities. The disincentives mentioned above may also weaken SNGs’ efforts to improve their tax administrations.

In most countries, intergovernmental transfers constitute the backbone of subnational finances, accounting for the bulk of revenues, especially at the intermediate level of government. They fulfil different objectives: filling vertical gaps, equalising revenue capacities and spending needs among SNGs, and funding subnational spending programmes regarded as national priorities or as having positive externalities. Different types of transfers privilege one or another of these objectives: revenue sharing is the most commonly used instrument to fill vertical imbalances; equalisation transfers aim to reduce horizontal imbalances; and block and special-purpose grants seek to finance subnational spending in sectors and programmes considered of national priority, or with large spillovers to other jurisdictions.

The relative weights of these objectives vary from country to country, reflecting a range of economic, institutional and socio-political factors. Accordingly, intergovernmental transfer systems differ significantly across countries. For example, shared revenues are very important in Argentina, Australia, Austria, Belgium, Colombia, Germany, Mexico and Spain, which are all characterised by large vertical imbalances. In other countries (e.g. Canada, India and the United States), block or special purpose transfers play more important roles. Such transfers are the primary source of subnational revenues in Indonesia, and a major one in other Asian countries, including Korea, Mongolia, the Philippines and Thailand. The degree of conditionality attached to such transfers varies widely across countries, types of programme and over time.

Most countries, including in Asia, have equalisation transfers, but the criteria for their distribution vary significantly. For instance, the Canadian equalisation transfer system only equalises revenue capacities, the South African one only spending needs, and the Australian one both revenue capacities and spending needs. Equalisation systems also vary widely in terms of the complexity of formulas and indicators used, reflecting particular countries’ preferences for transparency, popular acceptability of systems and the availability of relevant data.

The design of transfers (in particular, equalisation transfers) should be incentive-compatible, i.e. it should not discourage recipient governments’ efforts to exploit their own-revenue potentials nor their efforts to efficiently deliver the public services for which they are responsible. For this purpose, their horizontal distribution formulas should be based on indicators of revenue capacity and of spending needs (calculated assuming an appropriate level of spending efficiency), rather than on current or recent levels of revenues and expenditures. Accurately estimating such capacities and needs is, however, a complex task, requiring massive data availability. Most countries, including Asian ones, therefore, use more easily computable proxy indicators or focus only on revenue capacities (Chatry and Vincent, 2019[17]). Some delegate this task to independent bodies of technical experts, like the Australian Grants Commission (Ahmad and Searle, 2006[18]).

Several flaws in the design of intergovernmental transfers can adversely affect subnational capacities.15 In particular:

  • Transfers that are insufficient to cover the gap between revenue capacities and spending needs create unfunded spending mandates, undermining the capacity of subnational governments to carry out their assigned spending responsibilities adequately.

  • Transfers with a high degree of discretion (or with complex and opaque distribution formulas) create unpredictability for subnational budgets, as well as weaken the subnational budget constraint.

  • Revenue-sharing systems based on highly cyclical revenues (such as royalties from natural resources) impart excessive volatility to SNGs’ revenues and thus hinder their capacity to deliver a reasonably stable stream of essential public services.

In practice, transfer systems frequently suffer from one or more such shortcomings, not least because their design or implementation is often shaped more by political economy considerations than by economic principles and fiscal soundness. They also tend to be the frequent object of reforms, and thus change over time, more than other aspects of intergovernmental fiscal systems. The roles that vertical and horizontal intergovernmental co-operation can play, and has played in many instances, in improving the design of intergovernmental transfers is discussed further below.

Weaknesses in subnational governments’ capacities for carrying out their spending responsibilities with adequate effectiveness and efficiency can have several roots, including a lack of clarity in spending assignments, excessive fragmentation of jurisdictions, significant flaws in subnational planning and budgeting systems, and human resource (HR) constraints. This section briefly discusses each of these shortcomings in turn.

Most spending functions (except a few typically reserved to the central government, such as defence, foreign policy, border controls and the conduct of monetary policy) are shared among the different levels of government, although the respective shares of each vary significantly across countries and over time.

A recent OECD study (Dougherty and Phillips, 2019[19]), based on a survey of member countries, suggests that the national accounts-based shares of the CG, intermediate, and local governments in general government spending are not good indicators of their respective spending responsibilities, because they do not measure the degree of autonomy of each level in deciding, planning and implementing the spending. A finer disaggregation of spending power into indicators of autonomy in policy, budget, input and output-related decisions shows a wide variation across OECD countries in major areas of spending, namely education, health and long-term care, transport, housing and social assistance (Figure 1.2).

A comparison of a composite index of such indicators with spending shares suggests that there is limited correlation between the two, and that in general, subnational spending shares are significantly higher than their actual spending power in most areas.

Such an overlap of decision powers in all the main spending functions not reserved for the CG can give rise to inter-jurisdictional conflicts unless the respective responsibilities are clearly articulated in the relevant legislative and regulatory frameworks and are adhered to in practice. There are many examples in countries worldwide of such conflicts, occasionally escalating up to the constitutional courts.

A de jure or de facto lack of clarity in respective spending responsibilities can also weaken the political accountability of both national and subnational officials for the level, quality, and efficiency of public services, and/or lead to duplication of spending and wasteful use of public resources.

A number of presentations at the Roundtable of the Network on Fiscal Relations in Asia highlighted the pervasiveness and costs of unclear or overlapping spending responsibilities in Asian countries.

Subnational governments’ capacity to efficiently deliver goods and services for which they are responsible can also be adversely affected by excessive fragmentation of jurisdictions, especially at the local level. Municipalities with fewer than 5 000 inhabitants account for more than half of all local governments in the OECD area as a whole, although that proportion varies significantly across countries (de Mello, 2019[20]). In Asia, there is a vast dispersion of municipal sizes both across and within countries, mainly reflecting population size, geography and the degree of urbanisation. The average size of municipalities is particularly large in the People’s Republic of China and Indonesia (around 500 000 inhabitants), Korea and Malaysia (over 200 000 inhabitants) and quite small in the Philippines and Mongolia (around 2 000 inhabitants).

Small jurisdictions cannot exploit the economies of scale that prevail in various types of public services, such as energy and water provision, waste collection and disposal, and hospital care. They are also typically less capable of funding and implementing technologically advanced public expenditure management systems. For these reasons, central governments have frequently provided incentives for the merging of small neighbouring municipalities or encouraged contractual arrangements for a joint provision of services (Ter-Minassian and de Mello, 2017[21]).

In the Asia-Pacific region, Australia, Japan, Korea and New Zealand have shown a trend towards the merging of small municipalities, while other countries, such as India, the Russian Federation and Kazakhstan have provided increased autonomy to small villages.

Even medium-sized neighbouring localities or regions can benefit from co-ordinating the provision of network infrastructures in such areas as:

  • water resource management, to facilitate agreement on: the building of waterways from water-rich regions to more arid ones; the efficient utilisation of scarce water resources among different users; and the avoidance of upstream pollution16

  • environmental management and conservation, given the substantial potential for regional and even national spillovers of local activities in this area

  • energy management, distribution and conservation17

  • inter-regional and intermodal transport infrastructures, to maximise synergies between transport projects undertaken at the local and regional level, to improve connectivity.

Co-ordination needs are especially significant in large metropolitan areas, which often include several municipalities. Indeed, most such areas have created joint institutions (including jointly owned enterprises in some cases) for the management of public transport, water and sanitation, and other major utilities in the area (Bahl, Linn and Wetzel, 2013[22]; de Mello and Lago-Peñas, 2013[23]). France and the United Kingdom provide good examples of inter-jurisdictional co-operation in metropolitan areas.18

There is significant evidence from the 2014 OECD survey cited in the section “Some general considerations” and from available public financial management (PFM) diagnostics in developing countries, including in Asian countries, that weaknesses in planning processes hinder the capacity of SNGs to choose and deliver effective spending programmes. Specifically:

  • Subnational multi-year plans do not exist or, when they do, frequently they are not based on objective evidence of local needs or realistic estimates of resource availability.

  • Subnational plans are not aligned with national ones, and effective mechanisms to reconcile differences are lacking, with adverse effects on subnational access to sustainable financing for investments.

  • They frequently do not exploit the potential for synergies with those of neighbouring SNGs (e.g. in network infrastructure).

  • Most SNGs lack medium-term expenditure frameworks (MTEFs)19 to guide the annual budget process.

Most intermediate and medium-to-large local governments follow orderly and regular procedures for budget preparation, submission to the relevant legislature and monitoring and reporting budget execution. The evidence is more mixed for smaller, particularly rural, communities.

However, even in large and medium SNGs, budgeting processes often suffer from significant weaknesses:

  • limited coverage of budgets (i.e. significant recourse to extra-budgetary accounts and operations)

  • poor quality of forecasts of revenues and expenditures, partly reflecting uncertainties on intergovernmental transfers and on policy decisions by the central government that affect subnational spending, such as changes in civil servants’ wages and employment regulations, or other regulatory measures

  • little or no analysis of risks affecting revenue projections and spending needs, as well as of contingent liabilities, e.g. from the realisation of guarantees to enterprises owned by the SNGs

  • inadequate monitoring and reporting of spending commitments and liquidations, and consequently of payment arrears

  • lack, or limited coverage, of subnational treasury single accounts

  • accounting rules and procedures not fully aligned with national and international standards

  • fiscal reporting that is less timely than desirable, and does not adequately cover the operations of extra-budgetary units and subnational state-owned enterprises

  • limited debt management capacities and systems.

Weaknesses are also endemic in the management of subnational investments, a fact that is of particular concern, given the fact that in many countries the latter accounts for more than half of total public investments. Such weaknesses include:

  • only limited use of cost-benefit analysis in the evaluation of proposed investments

  • frequent and substantial delays and cost overruns in project execution

  • inadequate budgeting of operation and maintenance for new and existing infrastructure

  • absence of systematic ex post evaluations of completed projects

  • lacking access to affordable, long-term financing through capital markets, for most SNGs

  • limited (or no) capacity to successfully manage public-private partnerships (PPPs).20

Central governments can support the development of subnational fiscal capacities21 in several ways. The specific mix of desirable actions can be expected, of course, to vary both across countries and over time, reflecting the institutional characteristics of the intergovernmental fiscal relations system, economic and social conditions (in particular the extent of regional disparities) and the balance of political powers among levels government. CGs can both help subnational governments address resource constraints on their capacity and create incentives for them to do so.

Specifically, ways to ease resource constraints on subnational capacity include:

  • Assigning SNGs appropriate sources of own revenues, along the lines discussed in the section “The main weaknesses in subnational revenue capacities”.

  • Setting a floor on the rates of taxes assigned to SNGs, to minimise the scope for a race to the bottom.22

  • Sharing relevant taxpayers’ information with subnational tax administrations and supporting the latter in adopting more effective procedures in the collection and enforcement of their assigned taxes.

  • Avoiding unfunded expenditure mandates.

  • Correcting regional disparities through well-designed equalisation transfers, to the extent allowed by its resources and by society’s tolerance for such disparities.

  • Modulating the pace of devolution of spending responsibilities to SNGs according to their capacities to fulfil those responsibilities. It should be recognised that the scope for asymmetric decentralisation is often constrained by legal provisions or even by SNGs’ reluctance to accept a differentiation in spending responsibilities. There are, however successful examples of such an approach in various countries.23

  • Adopting appropriate reforms to increase flexibility and improve incentives for the civil service, and encourage (or if feasible, require) SNGs to do the same.

  • Supporting through technical assistance, and to the extent possible financially, improvements in subnational PFM systems.24

The central government can also create positive incentives for subnational capacity development through:

  • A clear assignment of expenditure responsibilities, not only in legislation but also in practice. This involves specifying, and adhering to, the respective roles in concurrent spending functions, and creating effective vertical co-ordination mechanisms, e.g. forums including representatives of the CG and those of the other level(s) of government responsible for those functions.

  • A design of equalisation transfers that does not discourage subnational own-revenue mobilisation efforts and rewards efficiency in spending, as discussed in the section “The main weaknesses in subnational revenue capacities”. The equalisation formula may even include financial rewards for above-average SNGs’ own-revenue efforts.

  • The development and use of appropriate indicators to monitor the effectiveness and efficiency of subnational spending programmes financed or co-financed through conditional special-purpose grants.

  • Well-designed and firmly enforced rules-based controls on subnational borrowing. The absence or ineffectiveness of such controls is a well-recognised source of subnational soft-budget constraints25 and a significant disincentive to both own-revenue-raising efforts and efficiency in spending.

  • Facilitating access to long-term financing for subnational investment projects, provided that they meet pre-specified quality standards.

  • Legislating minimum subnational PFM standards, if necessary, differentiated by the size of SNGs. This is especially important in the areas of accounting and reporting of subnational fiscal operations. The absence of such requirements, by preventing timely and reliable monitoring of subnational finances can give rise to significant fiscal risks, reduce the political accountability of subnational officials, and open space for corruption. A number of countries worldwide have moved towards the adoption of uniform public accounting standards and fiscal transparency practices in recent years (Irwin and Moretti, 2020[24]).26

As increasingly witnessed by international experiences, co-operation among subnational governments within each level of government can have beneficial effects on their respective capacities.

Horizontal co-operation forums can take different forms: they can involve SNGs’ highest level of leadership (governors or state premiers and mayors), sectorial authorities (e.g. state ministers, state or city secretaries), or technical officials; they can have broad or narrower remits; and their decisions can be more-or-less binding.27

Such forums can help shape common positions that strengthen the “voice” of the relevant level of government in its dialogue with the central government, e.g. in vertical co-operation forums and can facilitate agreement on important policy and institutional reforms.28

Horizontal co-operation can support the development of subnational fiscal capacities in several ways:

  • By helping to reduce adverse spillovers from subnational actions, e.g. predatory competition through taxes and subsidies, effluent discharges, inadequate control of contagious diseases.

  • By facilitating the exploitation of synergies in the delivery of services with economies of scale, and in network infrastructures, as discussed in “The main weaknesses in subnational spending capacities”.

  • Through the exchange of good practices, lessons from successes and failures, and positive demonstration effects.

  • By promoting an exchange of information and know-how among subnational revenue administrations. A good example in this area is the forum for co-operation among state revenue administrations in Brazil, which was instrumental in the development and adoption of the electronic VAT invoices in Brazil.

There is widespread consensus, both in the literature and among policy makers, that subnational governments’ capacities to effectively and efficiently fund and deliver the public goods and services for which they are responsible are crucial to realise the potential benefits of decentralisation. Although quantitative evidence of the state of such capacities across and within countries, as well as over time, is limited, there are sufficient qualitative bases for asserting that the vast majority of SNGs worldwide suffer from significant weaknesses in this respect.

This chapter has provided an overview of such weaknesses, ranging from inadequate assignments of own revenues, to flaws in tax administration, in the design of intergovernmental transfers, in spending assignments, and in various aspects of public financial management. The chapter has also discussed how central governments can contribute to easing the resource constraints on subnational capacities and can create appropriate incentives for SNGs to improve such capacities. Finally, it has briefly reviewed how intergovernmental co-operation can contribute to the development of subnational fiscal capacities.

There are no one-size-fits-all prescriptions in this area. Both the mix and the sequencing of actions by CGs and SNGs to improve the latter’s fiscal capacities need to be chosen taking into account the economic, social, and political realities of each country, as well as the more or less binding nature of legal constraints on those actions, and of the fiscal space to accommodate them. Successful international experiences with subnational capacity building can point to promising approaches, but a careful analysis of their applicability to different contexts is needed before they are embraced.

The first step in designing a strategy for capacity-building reforms in any individual SNG should be a diagnostic of the main roots of the existing weaknesses. This should be followed by the identification of the likely stakeholders involved in the correction of those weaknesses, i.e. the winners and the losers from a reform effort, as well as the identification of the institutional powers (e.g. the CG’s Executive or Parliament, or subnational own institutions) whose consent would be needed for the reforms. This would facilitate a realistic assessment of the political and institutional obstacles to the reform effort, and of possible approaches to confronting them.

The design of the specific components of the reform, and their timetable and sequencing, should be based on conservative assumptions regarding both the human and financial resources needed to carry them out, and the extent and timing of their effects. A well-tailored blend of actions with relatively short payoffs and more fundamental reforms with longer-term benefits can be expected to sustain the reform momentum more than alternative strategies.

Development partners (both multilateral and bilateral) can support subnational capacity development efforts by providing the benefit of cross-country experiences, and by augmenting national resources for the efforts, but the reforms must be nationally “owned”, both by the recipient subnational governments and by the central government when involved.


[9] Ahmad, E. and G. Brosio (eds.) (2015), Promoting Responsible and Sustainable Fiscal Decentralization, Edgar Elgar, Cheltenham.

[25] Ahmad, E. and G. Brosio (eds.) (2006), Handbook of Fiscal Federalism, Edward Elgar, Cheltenham.

[10] Ahmad, E. and G. Brosio (eds.) (2006), Normative versus Positive Theories of Revenue Assignments in Federations.

[18] Ahmad, E. and G. Brosio (eds.) (2006), On the Implementation of Transfers to Subnational Governments, Edward Elgar, Cheltenham.

[13] Artana, D. et al. (2012), “Sub-national Revenue Mobilization in Latin American and Caribbean Countries: The Case of Argentina”, IDB Working Paper, No. 297.

[22] Bahl, R., J. Linn and D. Wetzel (eds.) (2013), Financing Metropolitan Governments in Developing Countries, Lincoln Institute of Land Policy, Cambridge, Mass.

[14] Bahl, R., J. Martinez-Vazquez and J. Youngman (eds.) (2010), Challenging the Conventional Wisdom on the Property Tax, Lincoln Institute of Land Policy, Cambridge, Mass.

[11] Bird, R. (2010), “Subnational Taxation in Developing Countries: A Review of the Literature”, World Bank Policy Research Working Paper, No. 5450, World Bank, Washington, DC.

[29] Bird, R. (2007), The VAT in Developing and Transitional Economies, Cambridge University Press, Cambridge.

[30] Boadway, R. and A. Shah (eds.) (2007), Intergovernmental Transfers: Principles and Practice, World Bank, Washington, DC.

[17] Chatry, I. and R. Vincent (2019), “A global view of sub-national governments in Asia: Structure and finance”, in Fiscal Decentralisation and Inclusive Growth in Asia, OECD Publishing, Paris,

[20] de Mello, L. (2019), Local Government Cooperation: Challenges, Trends and International Experience.

[16] Detter, D. and S. Fölster (2017), The Public Wealth of Cities, Palgrave Macmillan.

[19] Dougherty, S. and L. Phillips (2019), “The spending power of sub-national decision makers across five policy sectors”, OECD Working Papers on Fiscal Federalism, No. 25, OECD Publishing, Paris,

[8] Eyraud, L. and L. Lusinyan (2011), “Decentralizing Spending More than Revenue: Does It Hurt Fiscal Performance?”, IMF Working Paper, No. 11/226.

[12] Hagemann, R. (2018), “Tax Policies for Inclusive Growth: Prescription versus Practice”, OECD Economic Policy Papers, No. 24, OECD Publishing, Paris,

[24] Irwin, T. and D. Moretti (2020), “Can subnational accounting give an early warning of fiscal risks?”, OECD Journal on Budgeting, Vol. 20/2,

[7] Kim, J. and S. Dougherty (eds.) (2019), Fiscal Decentralisation and Inclusive Growth in Asia, OECD Fiscal Federalism Studies, OECD Publishing, Paris/Korea Institute of Public Finance, Seoul,

[26] Kim, J. and S. Dougherty (eds.) (2018), Fiscal Decentralisation and Inclusive Growth, OECD Fiscal Federalism Studies, OECD Publishing, Paris/Korea Institute of Public Finance, Seoul,

[23] Lago-Peñas, S. and J. Martinez-Vazquez (eds.) (2013), Local Government Cooperation for Joint Provision: The Experiences of Brazil and Spain with Intermunicipal Consortia, Edgar Elgar, Cheltenham.

[27] Mizell, L. and D. Allain-Dupré (2013), “Creating Conditions for Effective Public Investment: Sub-national Capacities in a Multi-level Governance Context”, OECD Regional Development Working Papers, No. 2013/4, OECD Publishing, Paris,

[2] Musgrave, R. (1959), Theory of Public Finance, McGraw-Hill, New York.

[5] Oates, W. (2005), “Towards a Second Generation Theory of Fiscal Federalism”, International Tax and Public Finance, Vol. 12/4, pp. 349–373.

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

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

[4] Qian, Y. and B. Weingast (1997), “Federalism as a Commitment to Preserving Market Incentives”, Journal ofEconomic Perspectives, Vol. 11/Fall, pp. 83-92.

[33] Reyes-Tagle, G. (ed.) (2018), Bringing PPPs into the Sunlight: Synergies Now and Pitfalls Later?, IDB, Washington, DC.

[15] Slack, E. and R. Bird (2014), “The Political Economy of Property Tax Reform”, OECD Working Papers on Fiscal Federalism, No. 18, OECD Publishing, Paris,

[28] Ter-Minassian, T. (2012), Reform Priorities for Subnational Revenues in Brazil.

[21] Ter-Minassian, T. and L. de Mello (2017), Intergovernmental Fiscal Cooperation: International Experiences and Possible Lessons for Brazil.

[1] Tiebout, C. (1956), “A Pure Theory of Local Expenditures”, The Journal of Political Economy, Vol. 64/5.

[6] Weingast, B. (2009), “Second Generation Fiscal Federalism: The Implications of Fiscal Incentives”, Journal of Urban Economics, Vol. 65/3.

[32] World Bank (2013), Beyond the Annual Budget: Global Experience with Medium-Term Expenditure Frameworks, World Bank, Washington, DC.


← 1. Fiscal decentralisation refers to the transfer of revenue-raising powers and expenditure responsibilities from the national to the subnational levels of governments, and typically follows political decentralisation, which refers to the devolution of political powers to locally elected subnational authorities.

← 2. Ahmad and Brosio (2006[25]) provide a good review of such theories.

← 3. See Kim and Dougherty (2018[26]) for a selection of papers examining the role of fiscal decentralisation – or centralisation – in addressing the twin objectives of economic growth and reduced inequality.

← 4. The Revenue Administration Fiscal Information Tool (RA-FIT) is a web-based data-gathering tool to establish baselines of current revenue administration performance to improve comparative study and benchmarking. The Tax Administration Diagnostic Assessment Tool (TADAT) is a diagnostic tool to help governments gauge the performance of their tax administrations and identify priorities for reform. It focuses on nine key areas of performance of tax administrations worldwide. Additional details can be found at:

← 5. Public Expenditure and Financial Accountability (PEFA) is a tool for assessing the status of public financial management (PFM). A PEFA assessment provides an analysis of the performance of country’s PFM system at a specific point in time, based on internationally comparable criteria. The PEFA methodology can be reapplied in successive assessments to track changes over time. For more details, see

← 6. The Public Investment Management Assessment (PIMA) tool helps countries evaluate the strength of their public investment management practices. The PIMA evaluates the design and effectiveness of 15 institutions that shape decision making at three key stages of the public investment cycle, namely the planning, selection and implementation of investment projects. For more details:

← 7. The OECD conducted one such survey in 2013, focusing on subnational capacities in public investment management in OECD countries. The main results, reported in Mizell and Allain-Dupré (2013[27]) are discussed in the section, “The main weaknesses in subnational spending capacities”.

← 8. Own revenues are defined here as those on which SNGs have autonomy in defining the base and/or the rate structure. Under this definition, revenues shared with higher levels of government are considered intergovernmental transfers, if the definition of their base and/or rate is a prerogative of those levels. Unfortunately, most comparative datasets on subnational revenues (including the recently released IMF dataset on fiscal decentralisation covering 75 countries) classify revenues shared on a formula basis as own revenues.

← 9. In contrast, a subnational corporate income tax (CIT) (or a surcharge on the national CIT) is generally not recommendable on efficiency grounds because it is exportable outside the taxing jurisdiction and can be used for predatory competition; it is highly sensitive to the cycle; its base tends to be concentrated in relatively few jurisdictions where most corporations are headquartered; and it entails substantial administration and compliance costs.

← 10. See Ter-Minassian (2012[28]) for a comprehensive discussion of the problems with a subnational VAT in Brazil.

← 11. Bird (2007[29]) provides a good analysis of the performance of the Canadian VAT.

← 12. Although the base and rate structure of the Indian GST are set for the whole country by a high-level committee of the federal and state ministers of finance (the GST Council), it is jointly administered by the federal government (the Union) and the states, who collect their respective portion of the tax for goods produced and consumed within each state. The tax on interstate transactions is collected by the Union and shared with the states on a destination basis.

← 13. However, as demonstrated by the impact of the global financial crisis of 2008-09 on local finances in a number of advanced countries, property tax revenues can be substantially affected by the build-up and subsequent collapse of real estate price bubbles.

← 14. An iconic example in this respect is provided by the so-called “tax war” among the states in Brazil, which has severely eroded the revenue of the state VAT in that country (Ter-Minassian, 2012[28]).

← 15. There is a vast literature on the desirable characteristics of intergovernmental transfer systems [e.g. Boadway and Shah (2007[30]); Ahmad and Brosio (2006[25])].

← 16. Notable examples in this area are the Australian Intergovernmental Agreement of 2004 on the National Water Initiative and one on the management of the Murray-Darling Basin of 2013. Other examples are the intergovernmental Pact for the Management of Water Resources in Brazil, and the Delta Program in the Netherlands.

← 17. Australia offers a good example in this area with its 2009 National Partnership Agreement on Energy Efficiency.

← 18. See OECD (2015[31]) for various other examples of intergovernmental co-operation in metropolitan areas.

← 19. See World Bank (2013[32]) for a discussion of global experiences with MTEFs, requisites for their successful implementation, and benefits in terms of fiscal performance.

← 20. Reyes-Tagle (2018[33]) provides a comprehensive discussion of the demanding conditions for a successful use of PPPs in public investments at all levels of government.

← 21. Many of the remarks in this section also apply to the support that intermediate governments can provide to local ones under their jurisdiction.

← 22. Uruguay provides an example of successful vertical intergovernmental fiscal cooperation in stemming predatory competition by municipalities in the motor vehicles tax.

← 23. For instance, in Colombia, the CG only devolves some spending responsibilities to municipalities when they have demonstrable capacity to assume them, according to various objective indicators.

← 24. A good example in this respect is provided by Brazil, where the CG has supported extensive modernisation of state and municipal revenue administration and PFM systems through multi-year programmes financed by the Inter-American Development Bank with guarantees by the national treasury.

← 25. Ter-Minassian (2015[9]) provides a comprehensive discussion of the causes and consequences of subnational soft-budget constraints.

← 26. For instance, the European Commission has been working since 2011 on developing European Public Accounting Standards (EPSAS), adapting the International Public Accounting Standards (IPSAS) to EU conditions. EPSAS are to be implemented by EU members by 2025. In the meantime, the Commission is encouraging member states to move towards implementing the existing IPSAS on a voluntary basis. Brazil pioneered the adoption of common accounting standards for all public sector entities in its 2000 Fiscal Responsibility Law (FRL). In subsequent years, the federal government devoted substantial efforts and resources to assisting states and municipalities in implementing the accounting and reporting requirements of the FRL.

← 27. Ter-Minassian and de Mello (2017[21]) present examples of horizontal co-operation forums in several advanced and emerging countries.

← 28. Australia provides a good example of the use of an intergovernmental high-level forum (the Council of Australian Governments, or COAG) to build consensus for a major tax reform (introduction of a VAT shared between the Commonwealth and the states). In Germany, proposed reforms of shared taxes are discussed in various co-operation forums, to minimise the risk of their being blocked by the Upper House of Parliament (the Bundesrat), whose members are designated by the states and which has veto power on reforms affecting the states. In Belgium, the IFC was instrumental in ensuring agreement on a substantial subnational tax reform in 2001 (the Lambermont Agreement). In Mexico, the 2013 proposed tax reform package was extensively discussed in two technical co-operation forums (the Reunión Nacional de Funcionarios Fiscales and the Comisión Permanente de Funcionarios Fiscales) to help secure the political consensus needed for its approval by the Congress.

Metadata, Legal and Rights

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

© OECD 2020

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