Logo The World Bank Group Contact Us • Help/FAQ • Index • Search 
single line tv line
  Home > Topics > Public Sector Governance > Decentralization & Subnational Regional Economics > Intergovernmental Fiscal Relations
Decentralization Topics
What, Why, & Where
Intergovernmental Fiscal Relations
Service Delivery
Potential Impacts
Social Sectors
Subnational Regional Economics Topics
Bank Operations
Tools & Data
Useful Links
  Site Tools
Intergovernmental Fiscal Relations


Financial responsibility is a core component of decentralization. If local governments and private organizations are to carry out decentralized functions effectively, they must have an adequate level of revenues –either raised locally or transferred from the central government– as well as the authority to make decisions about expenditures. Fiscal decentralization can take many forms, including a) self-financing or cost recovery through user charges, b) co-financing or co-production arrangements through which the users participate in providing services and infrastructure through monetary or labor contributions; c) expansion of local revenues through property or sales taxes, or indirect charges; d) intergovernmental transfers that shift general revenues from taxes collected by the central government to local governments for general or specific uses; and e) authorization of municipal borrowing and the mobilization of either national or local government resources through loan guarantees. In many developing countries local governments or administrative units possess the legal authority to impose taxes, but the tax base is so weak and the dependence on central government subsidies so ingrained that no attempt is made to exercise that authority.

Expenditure Assignment

Fiscal decentralization involves shifting some responsibilities for expenditures and/or revenues to lower levels of government. One important factor in determining the type of fiscal decentralization is the extent to which subnational entities are given autonomy to determine the allocation of their expenditures. (The other important factor is their ability to raise revenue.) This note outlines principles and best practice and highlights how country specifics will ultimately be the best determinant of expenditure assignments.

It is important to clarify where local governments can determine the allocation of expenditures themselves versus those where the center mandates expenditures and local levels simply execute those expenditures. When statistics on subnational finance are available (most notably in the IMF’s Government Finance Statistics) these two types of expenditures are usually aggregated and one figure is provided on "percentage of subnational expenditures." Thus, analysts must be very careful when using this as in indicator of local autonomy, as autonomy will not be enhanced unless the funds are not tied by the center. (Of course, some countries do derive the benefits of local expenditure autonomy as a result of local governments’ ability to access the funds and circumvent the central mandates!)

Unitary and federal governments provide different opportunities for fiscal decentralization. Unitary countries do not have sub-national governments that are constitutionally empowered to make decisions over a specified range of government functions and services; rather, they have multiple subordinate levels of the same government (e.g., central, provincial, district). Federal governments, on the other hand, have constitutionally protected sub-national governments and thus, the possibilities for independent decision making are clearly stronger under these systems. It is important to note that local governments (as opposed to state or province) may not necessarily enjoy constitutional protection under federal systems. In practice, however, the extent and nature of decision-making power exercised by lower tiers varies widely from country to country in both federal and unitary countries and may change (in either direction) from time to time. At the extremes, some nominally federal countries (e.g., Venezuela) may be considerably less decentralized in reality than in nominally unitary countries (e.g., Colombia).

Despite the complexity of the existing situation in many countries, both theory and experience suggest strongly that it is important to state expenditure responsibilities as clearly as possible in order to enhance accountability and reduce unproductive overlap, duplication of authority, and legal challenges. Many would argue that decision-making should occur according to the principal of "subsidiarity" -- that is at the lowest level of government consistent with allocative efficiency (e.g., the geographic area that internalizes the benefits and costs of decision-making for a particular public service). The optimal size of jurisdiction for each service could theoretically differ, but in practice economies of administration and transactions costs lead to "grouping" of roughly congruent services at local (e.g., street lighting, refuse removal), regional (rural-urban roads, refuse disposal), and national (intercity highways, environmental policy) levels. Decentralized decision-making enlarges possibilities for local participation in development. In addition, national allocative objectives may be carried by local governments responding to incentives created by national grants and regulations as well as interlocal or interregional agreements. National governments have obvious roles with respect to both stabilization and distribution, and due attention must be paid to possible local conflicts with these policies.

In most cases, it is desirable that the national government assume responsibility for national public services, international affairs, monetary policy, regulation, transfers to persons and businesses, fiscal policy coordination, regional equity, redistribution (in which all levels of government may play a role), and preservation of an internal common market. However, some central functions such as regulation of the financial sector, environment, etc., may be effectively shared with sub-national governments. State governments may have dominant responsibility for education, health, social insurance, intermunicipal issues, and oversight of local governments. All local services should be assigned to local governments. In areas of shared responsibilities, the roles should be clarified. Generally, the central government should be involved with overall policy, setting standards, and auditing; state governments should have an oversight function; and local governments should be involved with the provision of infrastructure and services.

Assignment of public services to local or regional governments can be based on various considerations such as economies of scale, economies of scope (appropriate bundling of public services to improve efficiency through information and coordination economies and enhanced accountability through voter participation and cost recovery), cost/benefit spillovers, proximity to beneficiaries, consumer preferences, and flexibility in budgetary choices on composition of public spending. Assignment of responsibilities to various local governments could be asymmetric based on population size, rural/urban classification and fiscal capacity criteria. Thus large cities may have responsibilities for some services which are provided directly by the center in other areas.

An illustrative representative assignment of expenditure responsibilities is depicted in Table 1, taking into account also the possible desirability in some instance of "decentralization" beyond formal government to civil society. An additional important consideration to be borne in mind is that accountability is often best promoted by establishing a clear and close linkage between the costs and benefits of public services, so that the overall amount of expenditure responsibility to be assigned to a particular level of government will ideally correspond to the amount of revenues that level has at its potential command.

This table provides a guideline of expenditure assignments based on the subsidiarity principle. In practice, it is unlikely that any country would or should follow precisely the division set out in the table. In many instances some functions should be shared between levels of government in the sense that higher levels of government may exercise a regulatory or policy role, while lower levels of government are responsible for service delivery. Even within service delivery though, there are aspects best conducted by different levels of government. For example, with regard to infrastructure, the technical specifications for bridge construction might come from a higher level of government, construction and maintenance at the local level. In health care, the center may continue to provide technical training for staff, procurement of pharmaceuticals to benefit from economies of scale and ensure quality, and fund public health services. Intermediate levels might conduct supervision of local level personnel, provide refresher training courses, and, together with the local level, decide the appropriate mix of curative services to offer as well as ensure adequate maintenance of facilities and satisfaction with the personnel.

In this connection, it should be emphasized that assigning responsibility for the provision of service to a specific level does not imply that the same government should be directly engaged in its production. Collection, transport, and treatment services in solid waste management, for example, can be assigned to different public and private entities depending on economies of scale, commercial viability, and externalities. Many other services can also be "unbundled." Production decision should result from evaluation of alternatives using efficiency and equity criteria.

Experience has clearly shown that effective decentralization requires complementary adaptations in institutional arrangements for intergovernmental coordination, planning, budgeting, financial reporting, and implementation. Such arrangements may encompass both specific rules (e.g., in the design of fiscal transfers) and provision for regular intergovernmental meetings and periodic reviews of intergovernmental arrangements. Detailed central control over local use of funds is seldom appropriate. Instead, what is needed is transparency and accountability to local constituencies supported by strengthened higher level monitoring and reporting of local fiscal performance.

Back to top

Revenue Assignment

Governments rely on a wide variety of tax instruments available for their revenue needs, such as direct, indirect, general, specific, business and individual taxes. The question addressed here is which types of taxes are most suitable for use by each level of government.

Principles of Tax Assignment

The assignment of taxes by jurisdiction depends partly on the mix of various taxes used in the country overall. In public finance theory, the issue of the ideal tax mix even in the unitary state has not been widely developed. Governments almost universally employ balanced tax systems which have the feature that different taxes apply to basically the same bases. For example, general sales taxes, payroll taxes, and income taxes have bases which overlap considerably. From the point of view of standard efficiency and equity, one should be able to make do with a single general tax base, yet no governments behave that way. The usual reason given for this is that administrative considerations play an important role. A mix of taxes keeps the rate on any tax low, thereby reducing the incentive to evade or avoid the tax. Furthermore, by using a mix of taxes, taxpayers who would otherwise be able to avoid taxation of one type are caught in the net of another, making the tax system fairer. The importance of the various taxes in the overall mix remains, however, a matter of judgment rather than something that can be deduced from the principles.

These same general considerations apply in the case of assigning taxes in a federal government system. Efficiency and equity arguments have to be tempered by administrative considerations, and the exact assignment depends upon informed judgment. We can, however, outline the economic principles that come into play in deciding which taxes to assign to lower levels of government. They are as follows:

Efficiency of the Internal Common Market

The internal common market will be functioning efficiently if all resources (labor, capital, goods, and services) are free to move from one region to another without impediments or distortions imposed by policy. Decentralized tax systems can interfere with the efficiency of the economic union in two ways. For one, the uncoordinated setting of taxes is likely to lead to distortions in markets for resources which are mobile across states, especially capital and tradable goods. This problem will be lessened considerably if state governments recognize that resources are mobile. However, if they do recognize that, they may engage in socially wasteful beggar-thy-neighbor policies to attract resources to their own states. If all jurisdictions engage in such policies, the end result will simply be inefficiently low taxes (or high subsidies) on mobile factors.

National Equity

The tax-transfer system in one of the main instruments for achieving redistributive equity. The argument for making equity a federal objective is simply that all persons ought to enter into society’s ‘social welfare function’ on an equal basis, and presumably the federal government is the only level that can ensure that residents in different regions are treated equitably. This may be tempered if states have different tastes for redistribution, or if centralized decisionmaking is not guided by normative criteria. To the extent that equity is viewed as being a federal policy objective, decentralized taxes can interfere with the achievement of those objectives. As with the efficiency case, uncoordinated state tax policies may unwittingly induce arbitrary differences in redistributive consequences for residents of different states. Also, given the mobility of labor and capital across the states, the states may engage in perverse redistributive policies using both taxes and transfers to attract high-income persons and repel low-income ones. Beggar-thy-neighbor redistributive policies are likely to be offsetting with respect to resource allocation, but will result in less redistribution than in their absence. (Of course, those who abhor redistribution through government will prefer decentralized policies for precisely the same reason.) This is obviously likely to be more of a problem for those taxes which are redistributive in nature, as well as for transfers.

Administrative Costs

The decentralization of revenue raising can also serve to increase the cots of collection and compliance, both for the public sector and for the private sector. There are fixed costs associated with collecting any tax which will have to be borne for each tax type that is used by the states. Taxpayers will also have to incur costs of compliance for all taxes levied. The possibilities for evasion and avoidance will increase with decentralization for some types of taxes. This will be true where the tax base is mobile, or where the tax base straddles more than one jurisdiction. In the latter case, there will need to be rules for allocating tax revenues among jurisdiction; in their absence, some tax bases may face either double taxation or not taxation at all. Auditing procedures may also be distorted for those tax bases which involve transactions across state boundaries.

Fiscal Need

To ensure accountability, revenue means should be matched as closely as possible to revenue needs. Thus tax instruments intended to further specific policy objectives should be assigned to the level of government having the responsibility for such a service. Thus progressive redistributive taxes, stabilization instruments, and resource rent taxes would be suitable for assignment to the national government; while tolls on intermunicipal roads are suitably assigned to state governments. In countries with a federal level VAT, it may be too cumbersome to have sub-national sales taxes. In such circumstances, the fiscal need criterion would suggest allowing subnational governments access to taxes which are traditionally regarded as more suitable for national administration, such as personal income taxes.

Reconciling the Difference: Making Subnational Governments Accountable, While Preserving Efficiency and Avoiding National Distortions.

The main problem with the tax assignment that emerges from the preceding prescriptions, as illustrated in the attached table, is that it generally does not provide sufficient revenues for lower-tier governments. In part for this reason, local and especially intermediate-level governments in many countries levy a variety of specific (excise) taxes on gambling, motor vehicles, and so on. Again, however, such levies seldom produce anything like the revenue needed to finance a significant part of major expenditures such as education and health that are often assigned to subnational governments.

Resolving the Problem

The first way is to supplement local revenues by intergovernmental fiscal transfers - without undesirably reducing local efforts to collect their own taxes. (see transfer section)

The second option is to permit subnational governments to levy their own broad-based taxes - as long as they burden local beneficiaries only. In principle, a retail sales tax or a tax on personal income would be possible. In practice, however, the only efficient, desirable broad-based subnational tax that seems feasible is likely to be a flat-rate surtax (often called "piggybacking") on a national personal income tax. Retail sales taxes are seldom feasible in the circumstances of developing or transitional countries.

Whether with respect to such a surtax, a local property tax, or local taxes in general, the critical elements required to ensure local accountability without efficiency costs are (1) to restrict local governments as much as possible from exporting taxes and (2) to permit them to set their own tax rates. For efficiency, it may be desirable to assess the base of a tax centrally and even to have it collected by the central government; but for accountability it is critical that the local authorities are responsible (perhaps within limits) for setting the tax rate.

The Transition Economies: Redefining Local Powers

A special situation exists in a few transitional countries, including those of NIS, China and Vietnam, where local governments have traditionally had a larger role in administering national tax bases. In these transition economies the central government may not have full control over its tax bases due to local administration of these. For example, in China and Russia, revenues were collected at the local level (via a tax contracting system in China) and shared upwards. This created local level incentives to make better collection efforts for taxes fully retained at that level and less effort for taxes that were largely transferred upwards. Local governments in these countries preferred to receive transfers in kind or contributions from own enterprises rather than collect higher corporate taxes which had to be shared with higher levels. Revenue sharing on a tax by tax basis led to highly varied levels of efficiency in tax administration.

Further, in a country with conflict among levels of government, subnational administration of national taxes is not advisable since the subnational entity can refuse to submit national taxes if it becomes disgruntled (e.g., Tatarstan in Russia). China has recently strengthened its central tax administration to collect revenues from central and shared taxes. Second, problems are also caused by overlapping, uncoordinated administration, especially for sales and excise taxes.

Conclusion: The Balance Between Enough and Too Much

Decentralization has the potential to reduce accountability by breaking the links between the levels of taxation and expenditure. Major expenditure responsibilities are being transferred to local governments in an effort to improve service delivery, but there are few high-revenue taxes which can be assigned to local governments without creating national economic distortions. Efficiency in tax administration suggests that local governments should levy taxes on immobile factors (e.g. property taxes) and fiscal need criteria suggest that they should also levy cost recovery user charges such as frontage taxes (tax per linear front foot of property), tolls on local roads and poll taxes. These tax revenues are unlikely to be sufficient in many localities, and thus, intergovernmental transfers are required to mitigate this imbalance. While taxation increases can create constituent pressure for good local performance, some grant designs can create central government pressure for local performance.

Back to top

Intergovernmental Transfers/Grants Design

Intergovernmental transfers are the dominant source of revenues for subnational governments in most developing countries. The design of these transfers is of critical importance for efficiency and equity of local service provision and fiscal health of subnational goverments.

Taxonomy of Grants

For the purpose of economic analysis, grants can be broadly classified into two categories: non-matching and selective matching.

Nonmatching transfers: Non-matching transfers may be either selective (conditional) or general (unconditional).

Selective non-matching transfers offer a given amount of funds without local matching, provided they are spent for a particular purpose. Such conditionality will ensure that the recipient government’s spending on the specified category will be at least equal to the amount of grant monies. If the recipient is already spending an amount equal to grant funds, some or all of the grant funds may be diverted to other uses. In theory, due to fungibility of funds, increase in expenditures on the specified category would only at the limit equal to grant funds; in practice it is possible that the lumpiness of investments in areas such as infrastructure may result in increases in expenditures exceeding grants.

If the non-matching grant is unconditional or general, no constraints are put on how it is spent and unlike conditional grants, no minimum expenditure in any area is expected. Since the grant can be spent on any combination of public goods or services or to provide tax relief to residents, general non-matching assistance does not modify relative prices and is the least stimulative of local spending.

In some empirical studies, it has been observed that the portion of these grants retained for greater local spending tends to exceed local government’s own revenue relative to residents’ income; that is grant money tends to stick where it first lands. This is referred to as the "flypaper effect." The implication is that for political, technical and bureaucratic reasons, grants to local governments tend to result in more local spending than if the same transfers were made directly to local residents.

Selective (Conditional) Matching Transfers: Selective matching grants or cost-sharing programs require that funds be spent for specific purposes and that the recipient match the funds to some degree. Such a subsidy/transfer has two effects: (a) income effect – the subsidy gives the community more resources, some of which may go to acquire more of the assisted service; (b) price or substitution effect: since the subsidy reduces the relative price of assisted service, the community acquires more for a given budget. Hence both effects stimulate expenditures on the assisted category. Such transfers can be open-ended (no limit on matching funds) or closed-ended. Matching transfers may distort local priorities and be considered inequitable in that richer jurisdictions can raise matching funds more easily. But the latter problem can be offset, if desired, by varying matching rates with jurisdictional wealth and the former may be the desired outcome when the transfer is intended to e.g. internalize spillovers or achieve overriding national policy objectives.

The Conceptual Impact of Conditional Grants

Economic rationale for transfers and implications for grant design

We can identify five broad economic arguments for central-state transfers each of which is based on either efficiency or equity, and each of which may apply to varying degrees in actual federal economies.

  • The fiscal gap
  • Fiscal inequity
  • Fiscal inefficiency
  • Interstate spillovers
  • Fiscal harmonization

To the extent that the central government is interested in redistribution as a goal, there is a national interest in redistribution that occurs via the provision of public services by the subnational governments. Expenditure harmonization can be accomplished by the use of (non-matching) conditional grants, provided the conditions reflect national efficiency and equity concerns, and where there is a financial penalty associated with failure to comply with any of the conditions. In choosing such policies there will always be a trade-off between uniformity, which may encourage the free flow of goods and factors, and decentralization which may encourage innovation, efficiency and accountability.

As Bahl and Linn (1992) show and as discussed earlier, the most appropriate form of a transfer depends in large part upon its objective. Regardless of the particular design, however, experience demonstrates that good intergovernmental transfer programs have certain characteristics in common:

Transfers are determined as objectively and openly as possible, ideally by some well-established formula. They are not subject to hidden political negotiation. The transfer system may be decided by the central government alone, by a quasi-independent expert body (e.g., a grants commission), or by some formal system of central-local committees.

They are relatively stable from year to year to permit rational subnational budgeting but at the same time sufficiently flexible to ensure that national stabilization objectives are not thwarted by subnational finances. One system that appears to achieve this dual objective is to set the total level of transfers as a fixed proportion of total central revenues, subject to renegotiation periodically (say, every 3-5 years).

The formula (or formulae, if there is more than one grant) are transparent, based on credible factors, and as simple as possible. Unduly complex formulae are most unlikely to prove either feasible or credible in developing countries because there are often serious disputes on fundamental issues such as regional population sizes.
If several of the objectives discussed earlier are applicable - for example, some degree of equalization is desired while at the same time there are clear national policy objectives, e.g., with respect to the provision of minimal standards of education and perhaps variable degrees of national support for certain local infrastructure activities - it will generally assist both clarity and effectiveness if separate transfers are targeted at each objective.

Revenue Sharing

Many countries attempt to achieve various of the objectives ascribed above to transfers through systems variously described as "tax sharing" or "revenue sharing." While there are a wide variety of such systems, most of them - perhaps most markedly in the transitional countries - suffer from several common problems. First, if they are partial, that is, do not apply to all national taxes but only to a subset of such taxes, they may bias national tax policy. Second, if - as is often the case - they share the revenues from origin-based (production) taxes to the jurisdictions from which the revenues are collected, they break the desirable link between benefits and costs at the local level and hence reduce accountability and the efficiency of decentralization. Third, since in such systems tax rates are invariably set by the central government, and in addition since the sharing rate is often applied uniformly throughout the country, once again the accountability link is broken and subnational governments have no incentive to ensure that the amount and pattern of their spending is efficient. In addition, if, as in some of the transitional countries, such taxes are collected by local governments and then supposedly shared with national governments - and in this case perhaps especially if the sharing rates are higher (more flows upwards) for richer areas - either an undesirable disincentive for collection effort is created or, more usually, the temptation to "cook the books" is likely to be overwhelming.

Practical guidance on the design of these transfers is summarized in the
Criteria for the design of intergovernmental fiscal arrangements.

Back to top

Should Borrowing by the Public Sector Be Decentralized?

The debt crisis of subnational governments in Brazil, the inflationary impact of subnational financing in Argentina, and city-level bankruptcies in the United States have often been used to illustrate the possible macroeconomic implications of decentralizing borrowing powers. The moral hazard problem—the proposition that access to financial markets by subnational governments may create unplanned fiscal liabilities for central government—is the core of the argument.

The academic literature and country experiences, however, do not suggest an a priori adverse link between decentralizing borrowing powers and the central government's ability to maintain fiscal discipline and macroeconomic stability. Rather, the key seems to be the design of fiscal decentralization, particularly the design of the regulatory framework under which borrowing powers are decentralized.

Why Is Access to Financial Markets Important?

There are three primary reasons why access to financial markets is considered important for subnational governments:

Financing capital expenditure: Subnational governments often have responsibility for public investments that are lumpy in nature. Financing such capital investment through increases in current taxes would be inefficient. In addition, because the benefits of such public investments often last several decades, equity considerations would suggest that future generations participate in the financing. Capital markets provide this intertemporal link.

Matching expenditure and tax flows: Within a particular fiscal year, expenditures incurred and tax intake may not be fully in synchrony. Access to financial markets offers an opportunity to smooth out such mismatches.

Fostering political accountability: The pricing of capital by markets may provide an independent mechanism for fostering political accountability. Markets may signal the poor performance of subnational governments through increases in interest rates or simply by blocking access.
What Are the Mechanisms for Accessing Capital Markets by Subnational Tiers?

At least four channels exist for access to capital markets by subnational governments:

  • Direct borrowing by central government and on-lending to subnational tiers
  • Through a public intermediary, a state-owned financial institution
  • Direct borrowing from capital markets
  • Through market decentralization of public services, where possible (see below).

In ascertaining the relative merits of each channel, policymakers will need to consider several factors. The channel selected should minimize or eliminate the inefficient political allocation of credit. Any implicit liability by an upper-tier government should be explicitly recognized and reduced or eliminated. And selection should be made with a view toward strengthening capital markets.

Additional complications may occur when multiple channels are used. For example, in South Africa, government has created a public infrastructure bank that operates alongside a well developed private capital market. In this case, the regulatory framework will need to address the extent to which a complementary relationship can be established between the public and private financial systems, whether competition between the two is necessary and whether fair competition can be enabled, and the extent to which one channel displaces the other.

Evolution of Local Capital Markets

Of particular interest in assessing the merits of different channels is the linkage between access to capital markets by subnational governments and the evolution of local capital markets. In countries where local capital markets are weak, central government borrowing abroad may be the primary mechanism through which subnational governments access long-term finance. The challenge for policymakers would be to ensure a transition strategy for phasing out central government on-lending as capital markets emerge. In countries with more developed capital markets, however, the use of public intermediaries or central government on-lending will require careful attention to avoid displacing private capital.

Market decentralization—involving the private sector in the financing and delivery of government services—corporatization, and the unbundling of the delivery of services may also increase access to financing. For example, organizing the delivery of water or electricity in a company structure—private or public corporation—may improve access to capital markets if such organizations have a greater ability to shield themselves from political interference than do subnational governments. In countries with limited local capital markets, market decentralization may enable access to international capital markets directly by the service provider thereby facilitating the deepening of capital markets.

In addition, in relatively more developed capital markets, the use of private credit rating agencies and bond insurance agencies offers a market-based mechanism to monitor and regulate subnational borrowing. But these institutions also need to be monitored and regulated by the public sector.

Finally, one lesson from worldwide experiences is clear: subnational public financial institutions are undesirable. The incentives for subnational governments to abuse their relationship with their own financial institutions at the expense of macroeconomic stability is too great. Argentina stands out as a classic example of this situation.

How Should the Regulatory Framework for Subnational Governments Be Designed?

A well-designed regulatory framework is necessary to ensure that the decentralization of borrowing does not provide perverse incentives for excessive lending by markets and excessive borrowing by subnational governments—excesses that may eventually end up as liabilities of central authorities.

Such a framework requires transparency, preferably through information systems with standardized accounting systems for subnational governments and better public information on their liabilities. However, more public information will not by itself curb moral hazard problems. Penalties would need to be associated with excessive borrowing. One method is to legislate debt thresholds and penalties for crossing them and to establish transparent mechanisms for enforcing public bankruptcies. Examples of the latter include the U.S. type of financial control boards or the New Zealand system of court-appointed receiverships.

Ensuring that subnational governments have access to funds of their own (tax bases and unconditional grants) that can be pledged into the market as collateral is also necessary to reduce moral hazard tendencies. Without such direct fiscal backing, markets might view any capital market borrowing by subnational governments as being implicitly backed by the central government. Having their own fiscal base is therefore an important prerequisite for subnational governments’ access to finance and for limiting moral hazard problems.

Market decentralization may also offer an interesting mechanism for reducing the moral hazard problem. Privatization of public infrastructure could expose the firms to private sector bankruptcy laws and allow other private companies to bid for the assets in case of financial difficulties—an option unavailable with public ownership. (To implement this option, however, governments will need to ensure multiple deliverers in any one sector.) Private participation also creates options for accessing financial equity as well as debt, creating an incentive for equity to monitor debt that is not available in public financing systems. In addition, unbundling services may reduce the size of the entities involved, thus avoiding the "too big to fail" syndrome that often motivates bail-outs.

Finally, the back-door channels to creating financial liabilities need to be monitored and, where possible, closed. In particular, legislation must ensure that dipping into pension funds or using subnational corporations to borrow on behalf of subnational governments is not permitted or is explicitly included in debt limits. In addition, balanced budget requirements for subnational governments may ensure that current accounts are balanced by the end of each fiscal year so that borrowing to match expenditures and revenue streams does not lead to the financing of current account deficits over time.

Ultimately, the combined use of information systems, access to subnational governments’ own fiscal base, public legislation, bankruptcy laws, and market decentralization offers an institutional setting, creates an incentive to stick to a hard budget constraint at all tiers of government, and permits borrowing to be decentralized.

Should Foreign Borrowing Be Allowed?

Direct access to international capital markets by subnational governments is further complicated by issues of general capital controls, capital account liberalization, and the nature of the foreign exchange regime adopted by the government. This overall context of the exchange rate regime should determine whether direct borrowing by subnational governments in international markets should be permitted. Because this decision is further influenced by the depth of local capital markets, a sequencing issue may arise as well. Sometimes it may be preferable to allow direct access to local capital markets prior to opening access to international markets. Given the limited consensus on this issue, there seems to be a bias toward not granting direct access to international markets by subnational governments. In any case, the issues raised in this note about the regulatory framework for decentralizing borrowing apply equally to local and foreign borrowing.

Back to top