Descentralization Briefing Notes (WBI, 2000, PP 1-14 y 83-99)
Descentralization Briefing Notes (WBI, 2000, PP 1-14 y 83-99)
Descentralization Briefing Notes (WBI, 2000, PP 1-14 y 83-99)
Briefing Notes
BANK
Edited by
Jennie Litvack and Jessica Seddon INSTITUTE
WBI Working Papers
in collaboration with
Contributors
Junaid Ahmad, Harry Blair, Talib Esmail, PREM network
James Ford, Bert Hoffman, Graham Kerr,
Elizabeth King, Riitta-Liissa Kolehmainen-
Aitken, Ernst Lutz, Keith McLean, Dennis
Rondinelli, David Sewell, Anwar Shah,
Giulio De Tommaso, and Dana Weist
1
What and Why
This chapter contains two notes that describe the different forms of decentraliza-
tion and the rationale for decentralization. They provide a broad overview to the
several types of decentralization that can occur across countries, and even within
the same country and sector. Distinguishing among different types of decentraliza-
tion facilitates the discussion of policy design, and particularly of impact. For ex-
ample, whether a country chooses to “deconcentrate,” “delegate,” or “devolve”
certain functions—and the impact of those decisions—will depend on the
policymakers’ objectives, as well as on many factors related to the political, admin-
istrative, and fiscal structure of that country. It is important to introduce consis-
tency in any discussion of decentralization to avoid comparing apples and oranges
and to ensure that lessons can be elicited where appropriate.
-1-
Decentralization Briefing Notes
What Is Decentralization?
Decentralization—the transfer of authority and responsibility for public functions
from the central government to subordinate or quasi-independent government or-
ganizations or the private sector—covers a broad range of concepts. Each type of
decentralization—political, administrative, fiscal, and market—has different char-
acteristics, policy implications, and conditions for success. All these factors need to
be carefully considered before deciding whether projects or programs should sup-
port reorganization of financial, administrative, or service delivery systems.
While distinguishing among the different types of decentralization is useful
for highlighting its many dimensions and the need for coordination, these con-
cepts overlap considerably. Political, administrative, fiscal, and market decen-
tralization can appear in different forms and combinations across countries, within
countries, and even within sectors. Precise definitions are less important than
ensuring a comprehensive approach.
Political Decentralization
Political decentralization aims to give citizens and their elected representatives more
power in public decisionmaking. It is often associated with pluralistic politics and
representative government, but it can also support democratization by giving citi-
zens or their representatives more influence in formulating and implementing poli-
cies. Advocates of political decentralization assume that decisions made with greater
participation will be better informed and more relevant to diverse interests in soci-
ety than those made only by national political authorities. The concept implies that
the selection of representatives from local electoral jurisdictions allows citizens to
better know their political representatives and allows elected officials to better know
the needs and desires of their constituents.
Political decentralization often requires constitutional or statutory reforms, de-
velopment of pluralistic political parties, strengthening of legislatures, creation of
local political units, and encouragement of effective public interest groups.
Administrative Decentralization
Administrative decentralization seeks to redistribute authority, responsibility, and
financial resources for providing public services among different levels of govern-
ment. It is the transfer of responsibility for planning, financing, and managing cer-
tain public functions from the central government and its agencies to field units of
government agencies, subordinate units or levels of government, semi-autonomous
public authorities or corporations, or areawide, regional, or functional authorities.
Administrative decentralization has three major forms—deconcentration, del-
egation, and devolution—each with different characteristics.
-2-
What and Why
opportunities for local input vary: deconcentration can merely shift responsibilities
from central government officials in the capital city to those working in regions,
provinces, or districts, or it can create strong field administration or local adminis-
trative capacity under the supervision of central government ministries.
Fiscal Decentralization
Financial responsibility is a core component of decentralization. If local govern-
ments and private organizations are to carry out decentralized functions effectively,
they must have adequate revenues—raised locally or transferred from the central
government—as well as the authority to make expenditure decisions. Fiscal decen-
tralization can take many forms, including:
-3-
Decentralization Briefing Notes
-4-
What and Why
district, and local programs more effectively, and can provide better opportunities
for local residents to participate in decisionmaking. Decentralization may lead to
more creative, innovative, and responsive programs by allowing local experimenta-
tion. It can also increase political stability and national unity by allowing citizens to
better control public programs at the local level.
But decentralization is not a panacea, and it does have potential disadvantages.
Decentralization may not always be efficient, especially for standardized, routine,
network-based services. It can result in the loss of economies of scale and of control
over scarce financial resources by the central government. Weak administrative or
technical capacity at local levels may result in services being delivered less effi-
ciently and effectively in some areas of the country. Administrative responsibilities
may be transferred to local levels without adequate financial resources, making
equitable distribution or provision of services more difficult. Decentralization can
sometimes make coordination of national policies more complex and may allow
functions to be captured by local elites. Also, distrust between public and private
sectors may undermine cooperation at the local level.
Project and program planners must be able to assess the strengths and weaknesses
of public and private sector organizations in performing different types of functions.
Before developing elaborate plans for decentralization, they must assess the lowest
organizational level of government that performs functions efficiently and effectively
and, for functions that do not have to be provided by government, the most appropri-
ate forms of privatization. Even program planners who do not see decentralization as
their primary motive must carefully analyze the types of decentralization already present
in a country in order to tailor policy plans to existing structures. The success of decen-
tralization also frequently depends on proper training for both national and local offi-
cials in decentralized administration. Technical assistance is often required for local
governments, private enterprises, and local nongovernmental groups in the planning,
financing, and management of decentralized functions.
Centralization and decentralization are not either-or conditions. In most coun-
tries an appropriate balance of centralization and decentralization is essential to
the effective and efficient functioning of government. Not all functions can or should
be financed and managed in a decentralized fashion. And even when national gov-
ernments decentralize responsibilities, they often retain important policy and su-
pervisory roles. They must create or maintain the enabling conditions that allow
local units of administration or nongovernmental organizations to take on more
responsibilities. Central ministries often have crucial roles in promoting and sus-
taining decentralization by developing appropriate and effective national policies
and regulations for decentralization and strengthening local institutional capacity
to assume responsibility for new functions.
-5-
Decentralization Briefing Notes
-6-
What and Why
A number of theoretical and practical issues complicate the view of whether de-
centralization is a good or undesirable economic strategy. At the macroeconomic level,
an important concern has been that decentralization may make stabilization policies
more difficult to implement and may even lead to destabilizing levels and composi-
tion of overall public expenditures and public debt. However, decentralized systems
can be designed to avoid destabilizing effects and to ensure correct incentives. Some
of the decentralization of the 1980s, for example, was actually an offloading of fiscal
imbalances by central governments to subnational governments. Under these cir-
cumstances, it is not surprising to see a strong association between decentralization
and fiscal imbalances at lower levels. Another macro level controversy for which
evidence is inconclusive is whether decentralization retards economic growth.
Concerns about equity—interjurisdictional and interpersonal—have been cen-
tral to the discussion of decentralization. Some jurisdictions are better endowed
with resources than others, perhaps because of size or location. In addition, histori-
cal circumstances (such as apartheid) may have created local disparities. Thus an
intergovernmental fiscal program may be designed to shift resources to disadvan-
taged areas to ensure that all citizens enjoy a minimum level of service, regardless
of location, or receive enhanced assistance to accelerate amelioration of deficits,
because of location. The allocation of poverty program grants to subnational levels
should be analyzed carefully within a particular country context since lack of trans-
parency, or inadequate specificity in transfer design, sometimes results in wealthier
areas receiving more resources than poorer areas.
It is usually argued that central governments are ultimately responsible for ensur-
ing interpersonal equity. Where local economies are intrinsically open and many re-
sources, especially key human resources, are mobile, only limited success should be
expected from jurisdictionally focused distributional programs. Still, local governments
can and do play very important roles in implementing central distributional programs
and in determining a host of tax, expenditure, and intralocality transfer schemes.
The specific services to be decentralized and the type of decentralization will
depend on economies of scale affecting technical efficiency and the degree of
spillover effects beyond jurisdictional boundaries. These issues need to be taken
into account in the design of a decentralized system. In practice, all services do not
need to be decentralized in the same way or to the same degree.
In an important economic sense, the market is the ultimate form of decentrali-
zation in that consumers can acquire a tailored product from a choice of suppliers.
The nature of most local public services limits the full play of this option and estab-
lishes a government role in ensuring the provision of these services, but it does not
automatically require the public sector to be responsible for the delivery of all ser-
vices. Where it is possible to structure competition either in the delivery of a ser-
vice or for the right to deliver the service, the evidence indicates that the service
will be delivered more efficiently. Although uncommon in practice, local govern-
ments have successfully competed for the right to provide certain local services. In
an array of local public services in any particular country, a mix of solutions from
deconcentration to managed competition or privatization is likely to coexist.
-7-
Decentralization Briefing Notes
• The decentralization framework must link, at the margin, local financing and
fiscal authority to the service provision responsibilities and functions of the
local government, so that local politicians can deliver on their promises and
bear the costs of their decisions.
• Local communities must be informed about the costs of services and delivery
options and the resource envelope and its sources, so that the decisions they
make are meaningful. Participatory budgeting, as used in Porto Alegre, Brazil,
is one way to create this condition.
• Communities need a mechanism for expressing their preferences in a way that
is binding on politicians, so that there is a credible incentive for people to par-
ticipate.
• There must be a system of accountability based on public and transparent infor-
mation that enables communities to monitor the performance of the local gov-
ernment effectively and to react appropriately to that performance, so that
politicians and local officials have an incentive to be responsive.
• The instruments of decentralization—the legal and institutional framework, the
structure of service delivery responsibilities, and the intergovernmental fiscal
system—must be designed to support the political objectives.
-8-
2
Project Design: Political, Fiscal, and
Administrative Decentralization
This chapter contains nine notes that address different aspects of political, fiscal,
and administrative decentralization. Although these issues are often not consid-
ered together, they are all important components of successful decentralization.
For example, the “right” fiscal arrangements absent the appropriate regulatory struc-
ture or information and monitoring can lead to very different outcomes. All the
issues presented in the notes of this chapter interact and thus all must be consid-
ered when designing decentralization.
The first two notes in this chapter pertain to political decentralization. The first
note focuses on the constitutional, legal, and regulatory framework for decentrali-
zation and the second on the participation of society in decentralization.
The parameters within which decentralization can be designed depend on the
constitutional structure of the country, in particular whether it is a federal or unitary
government. In practice, the extent of autonomy provided to subnational levels var-
ies greatly; some federal countries resemble unitary countries (for example, Mexico),
and other unitary countries, de facto, operate as federal countries (China). Neverthe-
less, the legal opportunities and constraints for decentralization must be kept in mind.
These issues are explored in this note and are relevant to all the notes in this volume.
The extent to which decentralization is able to achieve benefits of allocative
efficiency depends in large part on the nature of local participation. Local leaders
can only reflect local needs and desires if channels for such input exist and if lead-
ers are accountable to their people. This note looks at the important relationship
between participation and decentralization.
The next four notes address important areas of fiscal federalism: expenditure
assignment, revenue assignment, intergovernmental transfer design, and
subnational borrowing. These four components make up the system of intergov-
ernmental finances. The impact of intergovernmental finances on the main eco-
nomic objectives of government—equity, efficiency, and macroeconomic
stability—depend on the system rather than any one component. For example,
significant decentralization of expenditures and revenues can lead to greater effi-
ciency and accountability in the wealthier parts of a country, but can lead to de-
clines in equity in the poorer parts unless intergovernmental transfers are used to
compensate the poorer areas. Similarly, expenditures and revenues need to be
matched when planning decentralization. If the center decentralizes more rev-
enues than expenditures, it can be starved of resources necessary to meet its na-
tional obligations; conversely, over decentralization of expenditures without
adequate revenues can lead to starving subnational levels of resources. The pres-
sure for unsustainable methods of finance increases in both these situations and
-9-
Decentralization Briefing Notes
- 10 -
Project Design: Political, Fiscal, and Administrative Decentralization
Substantially greater detail and specificity accompany the move down this three-
platform architecture from constitution to regulations, whereas greater difficulty
and a higher degree of authority (minister, parliament, and constitutional assem-
bly) are required to move up from regulations to the constitution. The levels of
difficulty and the locus of authority in effecting changes are important factors in
determining where in the architecture particular aspects of the decentralization
system are defined, as well as the relative specificity of the definitions. The rigidi-
ties and flexibility established in this structure have important implications for the
management of a decentralized system.
The placement of an item within this architecture may be the result of consen-
sus, but it is often the outcome of sometimes difficult negotiations between com-
peting interests. Those concerned with macroeconomic stability, for example, may
want intergovernmental fiscal rules to be a matter for regulation under the Minis-
ter of Finance, to give that ministry maximum flexibility in public expenditure
management. Local government advocates, by contrast, may argue (as they did
successfully in Brazil) for these fiscal distributional rules to be enshrined in the
constitution. In Uganda, the purposes and mechanisms for transfers are specified
in the constitution along with a formula for determining the minimum size of the
pool from which block grants are to be distributed; details of the distributional
formulas are the subject of regulations.
Because decentralization is a complex social experiment, a good case may be
made for allowing more flexibility to change the specificity of implementation in-
struments, while enshrining the political and philosophical principles in the con-
stitution and the operating structures in the laws.
- 11 -
Decentralization Briefing Notes
In addition to substantive law, a country’s procedural laws can have profound ef-
fects on the success of decentralization efforts. For example, when local expenditures
must be pre-audited by a central authority, rigidities are introduced that make the
benefits of decentralization more difficult to achieve. When reviewing the legal frame-
work for decentralization, it is not sufficient to examine laws specifically dealing
with decentralization. Other laws that mandate aspects of service delivery, civil ser-
vice, budgeting, and so on must also be considered to ensure a consistent approach.
Treatment of key issues in the legal and regulatory framework will be shaped by
the government’s structure as a unitary or a federal system. In some federal systems
(for example, Canada and India) local governments are completely under the au-
thority of state or provincial governments. The federal government is thereby lim-
ited in the relationships it may establish with the local level and must seek to affect
local behavior and outcomes through the states or provinces. A decentralization policy
such as India is trying to establish is significantly complicated by this factor.
Some unitary systems may exercise extremely centralized control over local
governments. In Indonesia, the Ministry of Home Affairs has had the authority to
appoint (and remove) mayors and even village heads. The structural impediments
in designing a decentralized system in this context are few, but that does not mean
that the process of instituting such a system is without critical hurdles. Indonesia
has had decentralization legislation on its books since 1974; the process there re-
mains far from completion.
As with other key aspects of decentralization, the legal and regulatory frame-
work will be tailored to country circumstances. Nevertheless, there are issues this
framework may be expected to address. Those issues of potential interest to the
work of the World Bank include the following:
• Classification of local governments within the tiers established under the con-
stitution
• Broad organizational structures and their roles and responsibilities
• Terms of office, operating powers, procedures, and limitations of the political
leadership, as distinct from the civil service
• The degree of autonomy of personnel policies and administration of local
governments
• The taxing and fiscal administration authority of local governments
• The borrowing authority and capacities of local governments
• The distribution of budgeting, expenditure management, accounting, auditing,
and reporting requirements
• Service provision and delivery authority
• The mechanisms for citizen participation and voice.
The legal and regulatory framework also should be designed to recognize dif-
ferences in management capacity. The assignment of functional responsibilities—
- 12 -
Project Design: Political, Fiscal, and Administrative Decentralization
Warning Flags
Five warning flags, selected from a potentially long list of downside risks, may
deserve special attention. First, local governments at the same nominal level may
vary considerably in their capacities. West Bank and Gaza, for example, have mu-
nicipalities that vary in population size from about 10,000 people to over 1 million,
with management capacities to match. Differences in fiscal capacity may be recog-
nized in the equity component of the intergovernmental fiscal system, but the fact
that management and administrative capacities also may vary substantially is rarely
accounted for. It is useful to have the legal and regulatory system recognize signifi-
cant differences in management capacities through a classification of local govern-
ment within levels. Policies and strategies to address these differences may then be
coherently considered.
Second, local governments should have the ability to borrow when they have
the capacity to repay. However, for moral hazard reasons (discussed in greater de-
tail in the section on borrowing) every effort must be made to promote the perspec-
tive that local government loans are internal obligations of local governments and
not of higher levels of government unless explicitly stated otherwise. The impor-
tance as well as the difficulties of doing this is illustrated by the circumstances of
subnational debt in Brazil. The legal and regulatory framework can support this
message by specifying the conditions under which local governments may borrow,
the limits of those borrowings, the reporting requirements for debt and debt ser-
vice, and the penalties for violating the rules.
Third, local government laws have not always anticipated the options, including
private participation and managed competition, that may be pursued in the delivery
of local public services. As a consequence, legal barriers may inappropriately restrain
the ability of local authorities to select the most desirable options for the delivery of
decentralized services. China’s cities have been imaginative in innovating and deliv-
ering some services not anticipated in the legal and regulatory framework within
which they operate; nevertheless, even under these circumstances rationalization is
desirable. Inappropriate barriers and constraints should be avoided or corrected in
the design and detailing of the legal and regulatory framework for decentralization.
Fourth, a voting democracy is often believed to satisfy the conditions for citi-
zen participation and voice in the design of decentralized systems, but in prac-
tice this may not be sufficient. Meaningful participation requires that citizens be
- 13 -
Decentralization Briefing Notes
informed and that their voices have impact where consequences are immediate.
At a minimum, the legal and regulatory system needs to provide for full, timely,
and easily accessible public disclosure of resource allocation decisions—in bud-
gets, procurement, and expenditure programs. An output and outcome orienta-
tion to expenditure management would be even more desirable. Uganda is
designing readily accessible budgets for all levels of government as part of an
expenditure management reform program that emphasizes an output and out-
come orientation. In addition, citizens must have reliable, secure access to the
means to enforce appropriate penalties for violations of rules.
Fifth, the terms of office for local political leaders are closely related to issues
of authority and accountability. Mayors need incentives to focus beyond the short
term to at least the medium term. This requires a long enough tenure or potential
tenure to be able to be seen to be accomplishing meaningful objectives. Mexico’s
three-year, nonrenewable mayoral terms, for example, have been associated with
a very short-term focus in local officials’ governance strategies. Where multiple
terms are allowed, three- to four-year terms are desirable. Where only single terms
are permitted, then five to six years would be appropriate. The detailed design of
authority, powers, accountability systems, and procedures must be related to lo-
cal circumstances, including issues that may range from cultural traditions to the
state of accounting and auditing systems. Considerations also include the bal-
ance to be struck between leaving room for aggressive leadership and protecting
the community against excesses, a choice that is a matter of political taste and is
often a consequence of historical experience.
The discrepancy between formal rules and actual practice regularly observed
in many countries is itself a commentary on the importance of certain character-
istics of the design and implementation of legal and regulatory systems. Ambi-
guity and complexity create openings for conflicting interpretation and confusion.
A single agreed interpretation is essential. Particular efforts to prepare and dis-
seminate popularized versions of the legal and regulatory system, as Uganda has
done, must be a key part of the decentralization strategy. Complexity is often
unavoidable, especially at the level of instruments for implementation. It helps,
however, if one instrument is not asked to do too much. This facilitates commu-
nication and implementation of the policy that the instrument is intended to sup-
port, better monitoring of the effectiveness of the instrument in that role, and
adjustment to the instrument or the policy.
- 14 -
4
Potential Impacts of Decentralization
This chapter contains four notes that review the existing evidence on the impact of
different forms of decentralization on equity, macroeconomic stability, economic
growth, and accountability, transparency, and corruption.
THE REALITY. The world does not fully correspond to this theoretical model, how-
ever. Many functions of subnational governments do, in practice, have significant
distributional effects: regulatory policies on land use and rent controls have pro-
found distributional implications, for example. Some frequently subnational func-
tions—most notably health and public education—also affect distribution.
Infrastructure services are another area in which local government policies
can have important redistributive effects, both progressive and regressive.
Throughout the world, infrastructure services are one of the core areas of respon-
sibility of local governments, an indication of the important role of this level of
1. Wallace E. Oates, Fiscal Federalism (New York: Harcourt Brace Jovanovich, 1972), pp.
6–8; R. A. Musgrave, The Theory of Public Finance (New York: McGraw-Hill, 1959).
- 83 -
Decentralization Briefing Notes
RECONCILING THEORY AND REALITY. Although debate on this subject continues, the
traditional view that central governments have the major role to play in functions
of income redistribution still prevails. Depending on how much a nation values
equity, the central government will need to play a redistributive role to raise wel-
fare for poor people in poor areas. Some of the most recent contributions to the
literature, however, recognize that local preferences for redistribution do matter.
2. Giancarlo Pola, George France, and Rosella Levaggi, eds., Developments in Local Gov-
ernment Finance: Theory and Policy (Cheltenham, U.K.: Elgar, 1996), contains several essays
on local governments’ revenues and responsbilities in Scandinavian countries.
3. David E. Wildasin, “Income Distribution in a Common Labor Market,” American
Economic Review 81(1991):757–74.
- 84 -
Potential Impacts of Decentralization
And while the traditional idea that the central government should control or at
least coordinate redistribution remains, there are some interesting new ideas for
approaches which incorporate subnational levels. In particular, Wildasin argues
that an optimal solution in multilevel systems of government is to provide grants
that, because of differing subnational preferences for redistribution, are nonuni-
form by subnational jurisdiction, but lead to identical levels of transfer payments.4
This researcher finds examples of behavior in multilevel systems of government
that support the analysis and policy prescription.
Factor mobility, the largest theoretical constraint on subnational redistribution
activities, may not be so large a concern in reality. First, while empirical studies
broadly indicate that interregional factor flows are influenced by the potential for
monetary gain, such flows are usually relatively small. Second, barriers to mobility
may be significant. For example, in many of the countries in transition, labor mo-
bility is limited because of lack of housing. Because housing is often the responsi-
bility of local governments in these countries, local redistributional policies (or the
absence thereof) acquire added importance.
VARIATION IN NET FISCAL BENEFITS. Variation in net fiscal benefits may arise for sev-
eral reasons. Two of the most common are:
- 85 -
Decentralization Briefing Notes
EQUITY ARGUMENTS. The equity problem for a central government’s tax and trans-
fer system arises because of the horizontal equity goal of treating citizens alike
no matter where they reside. The problem is that central government attempts to
attain interpersonal equity are implemented through tax and transfer systems
that deal only with the market or taxable incomes of individuals, but the net fis-
cal benefits from subnational government activity may lead to substantial differ-
ences in the real or comprehensive incomes of citizens in different areas. It is
difficult to think of practical ways for a central government to deal with the fiscal
benefits from subnational governments by adjusting its own personal taxes and
transfers. It would be impractical and politically unacceptable to have different
national personal income tax rates in different subnational jurisdictions, for in-
stance, and such a policy would also override legitimate subnational redistribu-
tion policies. One practical solution is to undertake intergovernmental transfers
to equalize subnational per capita fiscal capacities so as to make it financially
possible to achieve horizontal equity goals.
EXTENT OF EQUALIZATION PROGRAMS. This reasoning may help to explain why pro-
grams to equalize subnational per capita fiscal capacities are common throughout
the world. Programs of equalizing state and provincial fiscal capacities may be
more the rule than the exception in industrial countries with federal systems; they
6. William G. Watson, “An Estimate of the Welfare Gain from Fiscal Equalization,” Ca-
nadian Journal of Economics 19(1986):298–308.
- 86 -
Potential Impacts of Decentralization
7. For essays on equalization grants as well as other forms of transfers, see Ehtisham
Ahmad, ed., Financing Decentralized Expenditures: An International Comparison of Grants, Stud-
ies in Fiscal Federalism and State-Local Finance (Cheltenham, U.K., and Lyme, N.H.: Elgar, 1997).
- 87 -
Decentralization Briefing Notes
This is not necessarily the case. This will depend on how much emphasis the
central and local governments place on redistribution, how the system of inter-
governmental finance is designed, particularly transfers (see the sections on ex-
penditure, revenue, and transfers) and the extent to which participation in local
decisionmaking is truly broad-based (see the section on participation). Although
evidence is scant, decentralization seems to be more “pro-poor” when under-
taken at the local rather than intermediate level of government (see the section on
safety nets). This permits funds to reach the poorer areas and encourages greater
local accountability.
- 88 -
Potential Impacts of Decentralization
Subnational Incentives
Compared with central governments, subnational governments have less incen-
tive to consider the macroeconomic impact of their policies. Such impacts tend to
leak away to other jurisdictions, giving subnational macroeconomic stability poli-
cies public goods characteristics. Central governments are therefore usually assigned
the task of maintaining stability, and they should have the tools to go with it, such
as control over monetary policy and at least some control over fiscal conditions.
That includes a share of revenues and expenditures sufficiently large and suffi-
ciently flexible to influence aggregate demand in a country. However, whether in-
tended or not, subnational policies can influence stability: even balanced-budget
spending increases by subnational governments can affect macroeconomic stabil-
ity, and subnational borrowing can become a major macroeconomic concern. Yet
because macroeconomic shocks can hit different jurisdictions differently, some
subnational influence on macroeconomic conditions is actually desirable. Finally,
centralizing all expenditures and revenues with a potential macroeconomic impact
may have excessive efficiency costs.
8. William Dillinger, “Brazil’s State Debt Crisis: Lessons Learned,” Economic Notes 14,
Country Department 1, Latin America and the Caribbean (Washington, D.C.: World Bank, 1997).
- 89 -
Decentralization Briefing Notes
EXPENDITURE ASSIGNMENTS. Although its discussion of this topic is limited, the pub-
lic finance literature usually recommends assigning sufficiently stable expenditures
to subnational governments. For instance, many analysts argue that social security
is better assigned to the national level because of its procyclical nature. See the
section on expenditure assignment.
Incentives, however, may moderate this standard conclusion. For instance, while
assigning unemployment compensation to central government may create an “au-
tomatic stabilizer,” it also takes away the incentive for subnational governments to
hold down unemployment expenditures by, for instance, following a proactive
employment policy. Thus the centralization of social security may actually increase
macroeconomic imbalances, while making local governments responsible for a part
of these highly cyclical expenditures may reduce instability.
• Grants should be designed to fill the ex ante gap between subnational revenues
and expenditures. Filling ex post gaps could give rise to excessive expenditures
at the local level, as central government will foot part of the bill.
• The size of the grant pool should be fairly independent of macroeconomic
conditions, to avoid the procyclical spending patterns noted earlier. Having a
broad revenue base that feeds the grant pool is therefore to be preferred over
single-tax-based grants.
• Central government needs some discretion to adjust the amount of grants trans-
ferred. Although this is often impossible for general grants mechanisms, which are
usually based in law, special-purpose grants offer more flexibility. Their expendi-
ture effects can even be leveraged, if the grants require subnational matching.
- 90 -
Potential Impacts of Decentralization
- 91 -
Decentralization Briefing Notes
- 92 -
Potential Impacts of Decentralization
9. See Tao Zhang and Heng-fu Zou, “Fiscal Decentralization, Public Spending, and
Economic Growth in China,” Journal of Public Economics 67:2(1997):221–40; H. Davoodi, D.
Xie, and H. Zou, “Fiscal Decentralization and Economic Growth in the United States,” 1995
(unpublished); T. Zhang and H. Zou, “Fiscal Decentralization and Economic Growth: A Cross-
Country Study,” Journal of Urban Economics 43(1996). A critical summary of the evidence is
available in J. Martinez-Vazquez and R. McNab, “Fiscal Decentralization, Economic Growth,
and Democratic Governance,” paper prepared for the U.S. Agency for International Devel-
opment Conference on Economic Growth and Democratic Governance, Washington, D.C.,
October 9–10, 1997.
- 93 -
Decentralization Briefing Notes
10. David Wildasin, “Externalities and Bailouts: Hard and Soft Budget Constraints in
Intergovernmental Fiscal Relations,” Policy Research Working Paper 1843 (Washington, D.C.:
World Bank Development Research Group, 1997), discusses the incentives operating at vari-
ous levels of government.
11. For further discussion of these themes and examples, see the note on macroeconomic
stability; Charles McClure, “Comment on Prud’homme,” World Bank Research Observer
10(1995):221–26; Remy Prud’homme, “On the Dangers of Decentralization,” World Bank Eco-
nomic Review 9(1995); David O. Sewell, “The Dangers of Decentralization According to
Prud’homme: Some Further Aspects,” World Bank Research Observer 11:1(1996):143–50; Vito
- 94 -
Potential Impacts of Decentralization
autonomy to growing central government deficits, but these cases may not be gen-
eralizable because of their unique system of using local governments to collect cen-
tral government revenue. Other, more typical subnational governments in Argentina
and Brazil have followed the destabilizing path of expanding expenditures beyond
revenues and then turning to the central government for bailouts or debt relief.12
Switzerland, on the other hand, has been both highly decentralized and highly
stable in the post-war period. Research on Canada has shown that increasing
subnational governments’ budgets may have actually had a stabilizing effect.
Subnational governments have coordinated spending and may be better able to
offset region-specific business cycles.
Tanzi, “Fiscal Federalism and Decentralization: A Review of Some Efficiency and Macroeco-
nomic Aspects,” in Proceedings of the Annual Bank Conference on Development Economics 1995,
(Washington, D.C.: World Bank, 1995).
12. William Dillinger, “Brazil’s State Debt Crisis: Lessons Learned,” Economic Notes 14,
Country Department 1, Latin America and the Caribbean (Washington, D.C.: World Bank,
1997).
13. This hypothesis is particularly associated with Roy Bahl and Johannes Linn, “Fiscal
Decentralization and Intergovernmental Transfers in Less Developed Countries,” Publius:
The Journal of Federalism 24(Winter 1994):1–19.
14. Ariel Fiszbein, “Emergence of Local Capacity: Lessons from Colombia,” World De-
velopment 25:7(1997):1029–43.
- 95 -
Decentralization Briefing Notes
Types of Accountability
Accountability comes in two dimensions: accountability of government work-
ers to elected officials and accountability of the elected officials to the citizens
who elect them.
GOVERNMENT WORKERS TO LOCAL OFFICIALS. The first type of accountability can prove
difficult to achieve since civil servants, particularly professionals in such fields as
health, education, and agriculture—the very sectors that are most commonly de-
centralized—often have considerable incentive to evade control by locally elected
officials. Such people generally have university training and sophisticated lifestyle
practices that are hard to maintain in small towns and villages, career ambitions
that transcend the local level, and goals for their children’s education that local
schools cannot meet. They may also well fear that quality standards for service
delivery will suffer if provision is localized. Finally, they often find opportunities
for corruption are greater if they are supervised by distant managers through long
chains of command than if they must report to superiors close at hand. For all these
reasons, they tend to prefer strongly maintaining ties with their parent ministries
in the central government and resisting decentralization initiatives. And under-
standably, their colleagues at the center have a parallel interest in maintaining these
ties, for they are much concerned about preserving national standards in service
delivery and often about avoiding opportunities for venality (many corruption
schemes provide for sharing ill-gotten gains upward through bureaucratic chan-
nels to the top). See the section on the civil service for more about this subject.
Given all these reasons both good and bad for opposition, it is scarcely surpris-
ing that decentralization initiatives so often run into heavy bureaucratic resistance
and that designers find themselves pressured to keep significant linkages between
the field and the central ministries, especially concerning such issues as postings,
- 96 -
Potential Impacts of Decentralization
promotions, and salaries. Needless to say, such ties tend to undercut the capacity of
elected officials to supervise government servants who are supposedly working
for them. Some decentralized governance systems (for example, Karnataka State in
India) appear to have worked through these problems to establish popular control
over the bureaucracy, but it has taken many years to do so.
• Political parties can be a powerful tool for accountability when they are estab-
lished and vigorous at the local level, as in many Latin American countries.
They have a built-in incentive to uncover and publicize wrongdoing by the
party in power and to continuously present an alternative set of public policies
to the voters.
• Civil society and its precursor social capital enable citizens to articulate their
reaction to local government and to lobby officials to be responsive. These rep-
resentations generally come through NGOs (though spontaneous protests can
also be considered civil society), which, like political parties, often have parent
organizations at the provincial or national level.
• If citizens are to hold their government accountable, they must be able to find
out what it is doing. At the immediate neighborhood level, word of mouth is
perhaps sufficient for transmitting such information, but at any higher level
some form of media becomes essential. In some countries, print media can per-
form this function, but generally their coverage is minimal outside larger popu-
lation centers. A feasible substitute in many settings is low-wattage AM radio,
which is highly local, cheap to operate, and can offer news and talk shows ad-
dressing local issues.
• Public meetings can be an effective mechanism for encouraging citizens to ex-
press their views and obliging public officials to answer them. The cabildos abiertos
held in many Latin American countries are a good example. In some settings,
such meetings may be little more than briefing sessions, but in others they can
be effective in getting public officials to defend their actions. Participatory bud-
geting processes are also a powerful channel for communicating citizen prefer-
ences to local decisionmakers.
• Formal redress procedures have been included as an accountability mechanism
in some decentralization initiatives. Bolivia probably has the most elaborate
instrument along these lines with its municipal vigilance committees, which
- 97 -
Decentralization Briefing Notes
are based on traditional local social structures, charged with monitoring elected
councils, and encouraged to file actionable complaints with higher levels if
needed. In other systems, formal recall procedures are available to citizens dis-
satisfied with their officials.
• Opinion surveys have generally been considered too complex and sophisticated
to use at the local level, but usable and affordable technologies are being devel-
oped in the Philippines, enabling local NGOs through polls to assess public
opinion about service provision.
15. James Manor, “Democratic Decentralization in Africa and Asia,” IDS Bulletin
26:2(1995):81–88.
- 98 -
Potential Impacts of Decentralization
will ensure that government servants are responsible to elected officials, and that
officials are in turn responsible to the public that elected them. In the process these
systems of accountability should increase the pressure for more transparent local
governance, in which corruption will be easier to bring to light and thus to curtail.
But just as it took many decades for such efforts to make much headway in the
industrial countries, so too quick results cannot be expected elsewhere.
- 99 -