Table of Contents
Plenary Session One: Setting
the Context
Chair
Hon. Jaberi Bidaandi Ssali, Minister of Local Government, Uganda
Presenters
Professor Paul Smoke, Robert Wagner Graduate School of Public
Service, New York
Professor Dele Olowu, Institute of Social Studies, The Hague,
Netherlands
Discussants
Mr Roger Shotton, Technical Coordinator, UNCDF
Mr. Mamadou Diouf, Director of Local Government, Senegal
Professor Walter Oyugi, University of Nairobi, Kenya
Paper Presentations
Overview
of Decentralization and Workshop Theme Linkages
Paul Smoke
Decentralization is a complex and often somewhat elusive phenomenon.
Many countries around the world have been attempting, for several
reasons and with various degrees of intention and success, to
create or strengthen local governments in recent years. The
purpose of this workshop is to take stock of the situation in
Africa, with a particular view towards understanding what is
known about making decentralization work effectively in different
situations. We are also hoping to identify decentralization
issues that require further analysis and to consider how African
policymakers and practitioners can build on existing regional
channels of information and support for decentralization and
local government reform.
The five themes of the workshop, Fiscal Decentralization,
Sectoral Decentralization, Decentralized Environmental Management,
Political and Institutional Structures and Processes, and Local
Economic Development and Poverty Alleviation, were selected
as critical elements of decentralization around which considerable
interest and bodies of specialized expertise have developed.
At the same time, we must recognize that these themes are artificial
in the sense that decentralization cannot be so neatly carved
into independent dimensions. In fact, there are significant
linkages among these themes, and a principal task of the workshop
is to more fully understand them and their relevance for decentralization
policy and implementation in Africa.
Given the complexity of the topic and the vast array of issues
involved, what can be said in this short paper is necessarily
a simplification. Nevertheless, I will try to comment on a set
of issues intended to help to set the tone for the presentation
of the workshop themes and the working groups.
I will briefly cover four topics:
- The main goals of decentralization;
- The relationship of the workshop themes to these goals and
the linkages among them;
- Some myths and misperceptions about decentralization that
we need to be aware of; and
- The main challenges for decentralization policy analysts
and practitioners.
Three Goals (Potential Advantages) of Decentralization
There are many potential advantages of decentralization. Most
of them can be broadly captured under the classifications of
improved efficiency; better governance; and improved equity.
- Improved Efficiency
Proponents of decentralization argue that people are not
the same everywhere within a country, they have varying
needs for the public services that improve the welfare of
local residents and promote local economic development.
Sub-national governments are closer to the people, have
good access to local information, and understand the local
context well. If so, they can better identify the mix and
level of services that their particular constituents need
than can higher-level governments, thus improving what economists
call allocative efficiency. There are of course a number
of caveats. Some of the services local people want may best
be provided at greater scale by higher level government,
and some local services may affect other jurisdictions and
should not be left to the control of a single local government.
- Improved Governance
If local constituents see that their interactions with
their local governments lead to decisions about local public
services that are more consistent with their wishes than
decisions made by those at higher levels, they will feel
better connected to local governments. Being able to influence
the course of public affairs in at least some small way
that directly affect them empowers people, giving them a
sense of control and autonomy that they did not have before.
Again, there are caveats. Local constituents should not
necessarily have everything that they want, and certain
important national goals may have to take precedence over
local needs. Governance and collective action are not purely
local.
- Improved Equity
If local governments are made aware of variations in local
circumstances by their constituents, local governments are
in the best position to more equitably distribute public
resources and target poverty within their own jurisdictions.
They are of course constrained by their internal resources,
so that redistribution from richer to poorer areas must
be the responsibility of central governments. In addition,
local governments will not necessarily choose to pursue
redistribution in their jurisdictions unless they are forced
to do so by a broadly inclusive local political process
or interventionist central government.
Linkages among Three Major Dimensions of Decentralization
Having outlined the main goals and benefits of decentralization,
we now briefly consider how three of the most fundamental dimensions
of decentralization: fiscal, political, and institutional (including
local and intergovernmental), individually and collectively
affect its success.
- Fiscal Decentralization
Many associate the term "fiscal" only with financial resources,
but economists use it in a much broader way. Fiscal decentralization
comprises the assignment of responsibilities, including
sectoral functions and environmental management, as well
as the assignment of own-source revenues to sub-national
governments. There is an almost universal deficiency of
sub-national own source revenues relative to assigned sub-national
expenditure requirements. Intergovernmental transfers play
a critical role in closing this fiscal gap, as well as in
alleviating interregional resource disparities. Clearly,
without properly defined expenditure authority and adequate
resources, political and institutional decentralization
would have little impact. Poorly defined roles and resource
deficiencies can cripple local governments and undermine
the incentives of local officials to perform effectively.
Similarly, if local people participate in civic decision-making
processes and see no result because local officials have
inadequate power and resources to deliver services, they
may become disillusioned and even cynical about local government.
- Political Decentralization
Although fiscal and institutional decentralization are
critical, they are unlikely to bring about the major goals
of decentralization unless there is adequate political reform.
Sub-national governments may be empowered with clear and
appropriate functions and resources, and they may have appropriate
institutional mechanisms and capacity to spend resources
effectively. But in the context discussed above, efficiency
is predicated on the ability of sub-national governments
to understand the needs and preferences of local people
better than the central government. Without a well-developed
and broadly inclusive local political process, this is impossible,
and fiscal and institutional decentralization will not meet
their intended goals. We must keep in mind that fiscal and
institutional decentralization imply a reduction in the
accountability of sub-national governments to the center.
If this is not replaced by a degree of accountability to
local people, local officials may become primarily accountable
to themselves.
- Institutional Decentralization
Institutional decentralization refers to the administrative
bodies, systems, and mechanisms of government, both local
arrangements and the intergovernmental mechanisms that help
to manage and support decentralization. It also includes
mechanisms and procedures that link formal government bodies
to other key local actors, for example, traditional authorities,
non-governmental organizations, private sector partners.
This is the critical institutional architecture on which
decentralization is built. Interaction among levels of government
must be managed in a smooth way that facilitates local service
delivery rather than hinders it. Similarly, local civil
servants must function with an organizational structure
and procedures that allow them to meet their obligations,
and this includes a good working relationship with elected
representatives of the local people. Without appropriately
designed and implemented structures and processes, as well
as adequate institutional capacity to manage and deliver
the political and fiscal functions of sub-national governments,
decentralization will fail.
This brief exposition is a simplification, a caricature of
the complex relationships among various aspects of decentralization,
but it establishes the critical importance of considering linkages
as we move forward. The presentations of the theme paper authors
and their discussants, as well as our upcoming deliberations
in working groups, will help us to consider these linkages in
a more robust way.
Three Popular Myths and Misperceptions about Decentralization
Having briefly outlined some of the main goals of decentralization
and simply sketched some key linkages among its various dimensions,
I turn to some popular myths and misperceptions about decentralization.
We live in an era when a majority of developing countries are
decentralizing to various degrees, so there is inevitably a
great deal of rhetoric and generalization. A few issues are
particularly worth commenting on.
- Decentralization as a Panacea
Drawing on some of the possible benefits of decentralization
outlined above, a popular perception has developed in some
circles that decentralization is unambiguously desirable
and can alleviate many problems of the public sector. While
we do know that decentralization can have positive effects,
there can also be negative effects, some of which were implied
in the previous discussion of linkages. We know, for example,
that too much or inappropriately designed decentralization
can undermine central macro-economic control and worsen
interregional income disparities. Still, many alleged dangers
of decentralization, such as the extent and size of local
government budget deficits and local government fiscal irresponsibility
in repaying debt, are not inherent flaws of decentralization.
Instead, they are the result of poor system design, procedural
weaknesses, and capacity problems that ought to be the target
of sensible decentralization and local government reform
programs.
To argue that such problems require the maintenance of
centralization is equivalent to supporting the continuation
of dictatorships in countries where people are poorly educated
and unfamiliar with democratic concepts. The reality is
that neither the positive nor the negative claims about
decentralization have been definitively proven: most of
the "evidence" is anecdotal or based on selective analysis.
The prudent view is to be neutral about decentralization.
It can work if properly designed and applied, but it is
not a panacea for public sector ills, and the same level
of decentralization is not uniformly desirable across, or
even necessarily within countries or sectors. In short,
the potential benefits of decentralization are far from
automatic. Only if an appropriate degree of decentralization
is selected and local governments are given adequate support,
political, administrative, technical and financial, can
some or many of the purported benefits of decentralization
be realized.
- Decentralization Prerequisites
Some recent policy literature about decentralization argues
that there are certain individual or collective prerequisites
to decentralization: such as strong enabling frameworks,
effective local political systems, substantial locally derived
resources, and strong local capacity. This is somewhat misleading.
Much has been written, for example, about the critical need
for an adequate enabling framework in the form of constitutional
recognition of local governments and strong laws outlining
their roles and responsibilities.
There are, however, examples from around the world in
which some local governments have performed reasonably well
in the absence of this requirement, as well as cases where
central governments have undermined decentralization even
when there was an apparently strong framework in place.
In terms of the broader list, if all of these requirements
were truly prerequisites, no developing country would ever
be able to decentralize. Part of the definition of a less-developed
country is precisely that there is a lack of institutional,
political, and fiscal development. While it is true that
these mechanisms must eventually be in place to sustain
effective decentralization it will take considerable time
in most cases to develop them.
Thus, these so called decentralization prerequisites can
be more productively characterized as the basic elements
of a sensible program for building decentralization. Approaching
the problem from this perspective, the key policy questions
become: Which elements are already partially in place to
build on? Are there already some existing local capacities
that can be productively used? What priority elements need
to be developed? What is the most appropriate sequence of
reforms that will lead towards collectively developing these
elements in a sustainable way?
- The Primacy of Political Will
One of the ubiquitous claims about decentralization is
that lack of political will is the greatest impediment to
progress and the principal requirement for success. Although
it is evident that political will, which may come from the
central government or be forced on it by the people in times
when pressures for democratization are great, is important
for decentralization to succeed, it is not sufficient by
itself. We have seen several cases in recent years in which
dramatically changing political situations have led to major
constitutional and legal reforms and genuine attempts to
decentralize. Brazil's post-military government constitution
gave a powerful role to decentralized governments, as did
South Africa' s progressive post-apartheid constitution.
When Ethiopia developed its new constitution after the
fall of the Mengistu regime, they created a strong federal
system, hoping that, after the secession of Eritrea, they
could keep the country together by relaxing central control
of the largely ethnic-based states and giving them substantial
resources.
Uganda under Museveni, hoping to develop democracy for
its long deprived people and to undermine armed anti-government
movements operating in parts of the country, passed a constitution
and a local government Act that involved dramatic shifts
of power and finances to local governments. Other heavily
centralized systems undergoing major political changes,
such as Indonesia and Cambodia, are moving in this same
direction, but through legislation rather than constitutional
reform. All of these cases had great political will to create
decentralized systems, and most have made progress. But
none has attained the system they imagined creating when
they passed their new constitutions and laws. A few have
even re-centralized certain powers after taking significant
actions to give local governments more resources and autonomy.
Why? They have often been trying to do too much too quickly.
Some overwhelmed the capacity of weak local governments.
Some created opportunities for poor use of resources by
freeing local officials from central control without developing
their accountability to local constituents. Some severely
tested the bureaucratic resolve of the central ministries
who were losing power and resources more rapidly than they
could accept.
The reasons for highlighting these myths and misperceptions
about decentralization will, I hope, become clearer as we turn
to the final set of comments, which focus on how to approach
decentralization.
Three Key Challenges for Approaching Decentralization
Making progress with decentralization requires that many challenges
be confronted. One key challenge is defining an intergovernmental
system that makes sense in the context of a particular country
.A second is to create mechanisms for coordinating the activities
of the multiple actors invariably involved in decentralization
and to ensure that linkages among the key dimensions of decentralization
will be built. A third is to develop an appropriate strategy
for implementing decentralization.
- Defining Decentralization in Context
As emphasized above, decentralization is a complex, multi-faceted
phenomenon that can have positive or negative effects. The
specific form decentralization should take will vary across
countries with different political, institutional, fiscal,
and cultural characteristics. The history and traditions
of a country will determine what makes sense, particularly
in the near-term but also in the future. Thus, it is critical
to take careful stock of what is and what is not feasible
from the beginning, and how to deal with the always delicate
balance between developing genuine local autonomy and maintaining
justifiable levels of central control. An appropriately
defined system will help to reduce the probability of making
serious mistakes that will undermine decentralization before
it even begins. Certainly one thing we hope to learn from
this workshop is how the varying contexts of different countries
around Africa influence the form a decentralized system
should most appropriately take in order to maximize the
probability of attaining some of the benefits of decentralization
while minimizing the negative effects.
- Coordinating Actors and Building Linkages
Many of the aspects of decentralization outlined above
and embodied in the theme papers are the responsibility
of different actors. Ministries of Finance or National Treasuries
tend to have primary or substantial responsibility for decentralization
of revenue sources and the development of intergovernmental
transfer systems. Various sectoral line ministries, for
example, health, education, agriculture, public works, environment,
water, tend to be in charge of or to have significant influence
over the devolution of responsibilities under their general
expertise. Ministries of Local Government, Home Affairs,
or Interior generally have responsibility for developing
sub-national institutional structures and managerial procedures.
Sometimes they also have responsibility for local electoral
processes, or these may be under the control of Election
Commissions or other central agencies. Rarely is any single
agency clearly in charge, and even if one institution nominally
has a coordinating role, it often lacks enforcement authority.
Thus, these key dimensions of decentralization that we
know must function in concert are often at the mercy of
different agencies that may not fully support decentralization
and/or may be competing with each other for control of the
decentralization agenda and access to central government
and international donor resources. Some countries have experimented
with various types of coordination mechanisms, such as decentralization
secretariats, with varying success. A similar problem exists
locally, where it is necessary to coordinate activities
of local government councils, their operating agencies,
central government field offices, traditional authorities,
and non-governmental organizations, which in newly decentralizing
countries may have more capacity than formal local governments.
We hope that we can learn from this workshop about your
ideas and experiences with coordinating the actors involved
in decentralization and building linkages among the components
of decentralization at the national, intergovernmental,
and local levels.
- Decentralization Strategy
This brings us to the final point. What most decentralization
efforts lack, even those which involve an appropriately
defined and well-coordinated system backed by strong political
will, is a pragmatic implementation strategy. Decentralization
is not a once-and-for-all act. It is a complex process that
involves fundamental changes in attitudes about the way
that the public sector works. All major actors involved
are typically starting from positions inconsistent with
decentralization. First, central government officials, both
political and appointed, are often accustomed to making
major decisions and controlling local authorities extensively.
Second, local officials are often comfortable with heavy
central subsidization and control, and they may not be accustomed
to feeling accountable to local residents. Finally, local
people are often unaccustomed to paying for services or
expecting much responsiveness from local governments. These
long-held attitudes and behavior patterns will require years
and patient consultation to change. Even highly developed
countries that have undertaken decentralization, such as
France, have suffered through a slow and sometimes difficult
process.
What are the elements of a good decentralization implementation
strategy? This is a topic for discussion during the course
of the workshop, but three ideas are suggested:
- A general vision and framework are needed. This means
developing a process for defining decentralization in
a way that is appropriate for a particular country, a
point highlighted above. This may involve some experimentation
with different approaches to see what works best. Equally
important and consistent with the discussion above is
the need to develop appropriate processes and mechanisms
for coordination and building linkages;
- Reformers must be careful not to do too much too quickly,
and they should phase steps in a logical way. It is not
necessary to immediately turn over all constitutionally
or legally mandated responsibilities to all local governments.
Reforms with the greatest possibility of working effectively
in a relatively short time-frame should be undertaken
first. This requires prioritizing reforms, focusing on
simple sectors, tasks, and revenue sources that don't
immediately threaten in a significant way the tolerance
of the central power base or overwhelm local capacity.
Strategically differentiating among local governments
on the basis of capacity and performance rather than size
or official classification can also foster success and
improve resource use. Some local governments are likely
to be relatively capable and can be given greater responsibility,
while others will require significant training and technical
assistance. In a few cases, the precise reform steps have
been negotiated with local governments, placing some responsibility
on them if they fail to perform. Collectively, these aspects
of an initially modest, gradual, strategic approach should
raise prospects for early success, creating a base on
which to build further momentum in the future;
- Information, education, and incentives for behavioral
change are critical. All actors must understand how the
situation is to evolve and what is expected of them at
each step. Central officials need to support and monitor
decentralization in a coordinated way. Local officials
must learn to work with elected representatives and other
local actors. Perhaps most critically, local people themselves
must learn how to hold local officials responsible for
meeting their needs more fully. Often, alienated local
residents have to be gradually convinced that their local
governments can and will respond to them, and this requires
concrete results in the form of improved performance.
If some services are improved, people will be more willing
to pay local taxes. I have seen many cases, for example,
in which property tax increases were negotiated with the
local community in return for better service delivery.
Participation is not required for all local government
decisions, but it is certainly appropriate for some, and
heavier participation early on may help to raise the interest
of citizens in public affairs and their confidence in
holding their local officials accountable.
Conclusion
In applying these broad principles, the linkages among the
several aspects of decentralization and the need to build various
capacities must be taken explicitly into account. Gradual, for
example, does not mean focusing exclusively on one dimension
of decentralization first and ignoring others, such as holding
elections in year one and then devolving powers and resources
later. This will raise public expectations that local governments
will be unable to deliver on. Rather, it is preferable to sequence
decentralization to incorporate the various dimensions discussed
earlier throughout the process. In early stages where local
institutions and capacity are weak, for example, modest funding
(central grants with some local contribution) to support small-scale
projects/activities with carefully targeted technical assistance
would be a logical starting point. Conditions could be attached
to receiving the central resources and technical assistance.
These might include, for example, following an inclusive participatory
planning process, beginning to keep local records in a specified
systematic form, and making a minimum local contribution to
construction, operation, and maintenance costs. If a local government
meets the conditions and successfully plans and implements the
project, it can move on to progressively more substantial and
rewarding activities requiring increasingly stringent conditions
in subsequent years. More advanced local governments meeting
certain conditions could be treated in a more sophisticated
way from the beginning. This may involve some experimentation
with different approaches to see what works best. Equally important
and consistent with the discussion above is the need to develop
appropriate processes and mechanisms for coordination and building
linkages; process, local governments will have concrete incentives
to gradually and simultaneously improve their fiscal, managerial
and political performance, thereby progressively earning their
autonomy and realizing some of the potential benefits of decentralization.
Attempts to publicize good local government performance and
cases where sanctions were imposed for poor performance may
help to stimulate improvements in other places.
Given space constraints, these comments obviously represent
a caricature of a very complex process, yet decentralization
is by no means as easy or straightforward as this abbreviated
exposition might suggest. In addition, I am not arguing that
this particular type of process will work in every case. Most
of the workshop theme papers warn against generalization and
highlight the importance of context, and I fully endorse these
views. Systems have evolved in different ways and are at different
stages today. Despite such differences, it is fair to state
that the general requirements for effective decentralization
are universal:
- A vision for what the decentralized system should be and
what it is expected to accomplish;
- Some basic framework that defines- in an adequate and enforceable
way-the key components of the system and their linkages;
- A carefully defined strategy for how to bring the system
into existence and to support its evolution over time.
This latter component is neglected almost everywhere, even
though it is in many ways the most critical element of effective
decentralization. The key challenge for African policymakers
is to search for informed and creative ways to define and implement
their vision of decentralization both gradually and strategically.
Local Political and Institutional
Structures and Processes
Professor Dele Olowu, Institute of Social Studies, The
Hague, Netherlands
African countries are at once regarded both as centralized
and excessively decentralized. The formal state structure is
centralized, but society is decentralized and people live in
scattered rural communities, although this is changing rapidly
as more people move to the continent's growing cities.
In the last few decades, some African countries have embarked
upon decentralization programs and policies that are different
from previous efforts in two main senses:
First, their primary objective is to empower the people
as a part of the effort at democratizing state institutions
and initiate/support local self-governing structures, not the
extension of state control. Indeed, some countries have construed
the building of the capacity for local democracy and working
upwards as a more viable strategy of state democratization rather
than the simultaneous democratization of all of state institutions.
Second, there is a growing appreciation of the need
to develop not local government as such, but local governance.
This implies not only the vertical transfer of responsibilities
and resources from central to local governments (the conventional
conception of democratic or devolutionary decentralization,
DD) but also the development of horizontal networks between
local governments and local non-state actors: the private sector,
civil society, and international organizations. Such networks
facilitate local governance in three ways. First, they help
to mobilize additional resources for local level development
that would have remained dormant. Secondly, and perhaps, more
importantly, they enhance the accountability of local level
political and management officials. As discussed below, accountability
of local governance institutions constitutes a major problem
of democratic decentralization programs. Thirdly, synergies
between these institutions lead to innovations and higher levels
of productivity as disconnected structures are networked into
a common framework at the local level. A number of countries
on the continent have continued with the old approaches to decentralization,
which aggravate central control over all local activities for
reasons that appear sound, for example, weak local capacity,
fear of local corruption, communal and ethnic insurrection.
However, this paper focuses on innovative experiments to highlight
how they have tackled the dilemmas that confront DD in countries
characterized by state weakness, grinding poverty and poorly
institutionalized state structures.
In particular, the paper focuses on two key questions:
- Has decentralization produced mechanisms to mediate relationships
between the central and local governments that support and
facilitate strong and effective local level governance? One
of the most important lessons of successful decentralization
programs is that DD involves the redefinition of central and
local government tasks and resources and neither of these
can be achieved without effective mechanisms for facilitating
intergovernmental relations. Effective intergovernmental relations
require institutional mechanisms that support the redefined
relationships between central and local organs to achieve
the objectives of DD. This is important given strong open
and subtle resistance to DD by central level actors;
- Has decentralization fostered the development of inter-local
organizational synergies at the community level? This is the
effort to build critical institutions for local governance
involving both state and non-state institutions. This last
issue is important because one of the strongest arguments
against DD is fear of creating local despotism, which further
undermine local development. This paper highlights a number
of mechanisms created by communities - in urban and rural
areas of Africa - to promote effective local governance. Particular
emphasis is given to rural forms.
To provide a context for the discussion, the paper begins
by reviewing the motivations for local government decentralization
in Africa. Next, it highlights the main dilemmas confronted
by DD programs in the region. The latter analysis will underscore
among others two political problems: reluctance to decentralize
and the possibility of local elite capture even when power is
decentralized.
Patterns of African Decentralization:
Comparisons of Past and Present
Compared to all other regions of the world, Africa has the
weakest formal local government structures, judged in terms
of size of expenditures and employees. Yet, decentralization
is not new to the region. Since the colonial period, diverse
attempts have been made to decentralize state structures. The
economic crisis of the 1970s, which was followed by structural
adjustment and political reforms in the 1980s and 1990s, led
to new initiatives in decentralization designed to improve state
systems that were widely regarded to have failed or collapsed.
By the end of the 1990s, it was possible to classify African
countries in terms of those with progressive policies on DD
and those that had stuck with the old forms of deconcentrating.
In a third category are those that have gone further to experiment
with federal arrangements. A sample of countries in these different
groups is provided below:
- Deconcentrating with nominal devolution: Ghana, Cameroon,
Kenya, Malawi, Zambia;
- Devolution: Mauritius, Senegal, Uganda, Sierra Leone, Mali,
Cote D'Ivoire, Nigeria, Tanzania;
- Partial Devolution (Urban Areas Only): Namibia, Mozambique,
Botswana;
- Federations: Republic of South Africa, Ethiopia, Nigeria,
and Tanzania.
Factors Motivating Democratic Decentralization (DD) in
Africa
The following factors help explain the renewed interest in
new programs and policies on democratic decentralization in
Africa from the 1980s:
- The failure of centralized public sector management evidenced
by economic, fiscal and political crises;
- The resulting decline in state resources increased pressure
for economic, institutional and political reforms as part
of the search for new paradigms of governance;
- Pressures for political reforms by domestic actors outside
the state, many of who had become more visible politically,
and more sophisticated in their critique and protest of state
policies as the economic crisis bit harder. State officials,
in response, sought institutional arrangements that could
bridge the wide gulf between the state and the citizen in
the policy processes and the synchronization of informal local
realities with the highly centralized structures and operations
of the (formal) state system;
- Pressure from external donors: an important consideration
given the fact many African states are heavily dependent on
donor funds for development expenditures;
- Urbanization is variable in Africa but it is marked everywhere
by high rates of growth. This phenomenon is itself also a
symptom of fundamental changes to the modes of economic production;
- The use of decentralization by ruling groups to neutralize
or seek compromises with regional or local elites. It played
a role for instance in the resolution of the long conflicts
in Ethiopia, South Africa and Mozambique and looks promising
in the resolution of raging conflicts in, for example, the
Sudan, Morocco, Angola, Senegal, the two Congos;
- The pressure for globalization compelled many national governments
to focus their attention on strategic issues of national economic
and political management. Detailed management of cities and
communities could be conveniently left to other institutional
actors - some of which, like the religious and philanthropic
institutions, have in the past demonstrated their capacity
to undertake local economic development.
Dilemmas of Democratic Decentralization in Africa
- Political Dilemmas
Devolutionary decentralization confronts two major political
problems:
- The first is the unwillingness of political and administrative
leaders to share monopoly power inherited from the colonial
period. This is in part due to fears that devolution might
undermine national cohesion and fan the embers of secession
in societies in which ethnic and community loyalties are
quite strong. There is also the fear that devolution might
compromise the integrity of nationally delivered services.
In many instances, these rational fears are often a cloak
for the fear on the part of ruling elites at the center
of loosening their grip on political power. Devolution
is viewed as a zero-sum power game in which local actors
gain at the expense of the center, rather than a positive-sum
power game in which all players, both at the local level
and at the center gain over time;
- A second political dilemma is the problem of local elite
capture. In many instances, it is local elites rather
than the most vulnerable that capture decentralized power
- which is then utilized to repress local minorities -
including women and other marginal groups.
These two political dilemmas must be overcome before any
country can have effective and institutionalized local governance
systems. The first problem has been overcome in a number
of countries but none can be said to have fully tackled
the second problem. Evidence exists that shows that some
countries have been able to forge a political coalition
to get DD adopted. Critical institutions include constitutional
re-arrangement of responsibilities and powers of central
and local governments, the constitutional transfer of funds
to local governments, and the transfer of the power to make
and approve own budgets and personnel, including those transferred
to the local governments from the national governments.
Elaborate systems of fiscal transfers have been put in place
in countries such as Nigeria, Uganda, and Ethiopia, which
ensure that substantial national resources are made available
to cash-starved local organs. Human resource management
systems have also been developed at the local level, making
them almost at par with the national level governments.
On the other hand, very few countries have been able to
tackle the second problem of developing effective local
government structures that promote accountability, public
private partnerships and effective services delivery at
the local level.
Economic and management dilemmas are secondary to this
paper but critically linked to political reform.
- Economic dilemmas
Economic dilemmas include:
- Most of the advantages of DD are based on their contributions
to the allocative functions of government, but decentralization
can complicate stabilization and redistribution, as DD
limits the ability of the center to control local spending;
- DD without wider institutional, political and economic
reforms is a contradiction, as the improvements in allocative
efficiency can only be realized if local governments are
accountable to citizens and have the capacity to perform;
- Many countries undertake decentralization to reduce
budgetary pressures, but decentralization in the early
stages often involves heavy development and capacity building
costs.
- Management dilemmas
Management dilemmas include the problem of appropriate
institutional design that respond adequately to the above-mentioned
problems and how to sustain the program of decentralization
once initiated. The problem is compounded by the absence
of reliable data on governmental performance and also by
weak capacity at central and local levels. This third set
of problems are usually the most difficult to resolve. Some
of the most difficult design and management issues include
how to make decentralization policies applicable to all
parts of a country, given their different levels of development-including
the possibilities of adopting a phased approach; and how
to transform the zero-sum game perception of decentralization
into a positive-sum game in which all sides gain.
Intergovernmental Relations Mechanisms for Effective Local
Governance
The first political challenge of DD is how to develop mechanisms
to ensure smooth intergovernmental relations in the allocation
of responsibilities, finance and human resources management
in selected countries. Some cases are discussed below.
Nigeria started DD reform early, but it was
also closely associated with the effort at political and economic
reforms in this troubled country. The military took over power
for the first time in 1966 and after staying in office for ten
years began to implement a program of democratization. Local
governments were perceived as one of the five key principles
for the restoration of civilian government in a way that would
arrest the abuses of the past. After a series of consultations
in 1976, a nation-wide program of local government reform was
announced. This program included all the principles of local
government or DD reforms enunciated above: autonomous local
governments with clearly articulated responsibilities and financial
resources. The Nigerian federal government felt that the LG
reforms would help to disperse the country's public sector investments
and to redistribute basic amenities to all parts of this huge
country. In this, they have been largely successful. Key mechanisms
highlighted in the full paper include:
- The constitution of 1979 and the country's subsequent constitutions
(1989, 1995, 1999) that progressively incorporated major decentralization
reforms;
- The emergence of a permanent body on fiscal revenue mobilization
and sharing among federal, state and local governments;
- The creation of special-purpose bodies (Directorate of Foods,
Roads and Rural Infrastructures; the Community Bank) to work
with local governments to promote rural development; and adoption
of a unified model for human resource management.
Ethiopia was ruled for many years first as a
feudal empire, and then as a highly centralized Leninist military
socialist state. Ethiopia adopted a new federal constitution
in 1995 which stipulates that 'every nationality and people...has
the right to a full measure of self-government which includes
the right to establish institutions of government in the territory
that it inhabits...'. The regions vary widely in population
and competence. They have wide (residual) powers, including
the ability to secede from the federation. They are empowered
'to establish a state administration that best advances self-rule
and a democratic order ' in their states. Their responsibilities
include the formulation and execution of economic and social
development policies, land and natural resource, state police
forces and maintenance of public order. Below the regions are
zones, woredas (or districts) and kebeles (communes) - all of
which are assigned specific responsibilities within the province
of the state but devolution is mainly felt at the level of the
woredas. They formulate their own budgets and have assigned
responsibilities.
The full paper reviews the mechanisms for finance, and human
resources management in the new Ethiopia federation and discusses
the challenges confronted by the resource-rich versus the poor-resource
based states, including how this affects the arrangement for
human resources management.
The full paper also discusses national mechanisms in Cote
D'Ivoire and Uganda for sharing responsibilities (the constitution
and the statute), financial resources (the fiscal commission)
and the management of human resources (the district service
commissions). It will also argue that DD programs require the
making of critical choices concerning the level of differences
they can allow between different parts of the same country.
Building Effective Inter-organizational Relations at the
Community Level
The second major political problem that DD efforts confront
is how to build effective organs for participation and accountability
at the local level. These must ensure that the expectations
of higher levels of public participation, transparency and accountability
are actually met at the local level and preclude local elite
capture. This represents a dilemma for the design of DD systems.
Local elites are required for the success of local government
systems: they bring resources, knowledge networks that make
these systems become fully operational and effective. At the
same time, they must not be allowed to exclude the masses of
the people; otherwise serious problems of equity, responsiveness
and corruption at the local level are raised.
The key is to create institutions that enhance public participation,
accountability and the transparency of local level institutions
to their constituents.
The question then is: what types of institutional mechanisms
help to effectively neutralize local elite capture? This issue
raises another dilemma. Many countries have tried to resolve
the problem of accountability of local governments by asserting
central supervision and accountability of local governments
to higher organs of government. Unfortunately, these have not
produced the best results either, partly because central governments
also have problems maintaining effective accountability regimes
at national and local levels. For instance, central officials
abuse the audit process in many countries. A combination of
mechanisms of central government (including field administration)
and local account-ability is often needed, but what constitutes
the balance between the two?
Available research shows that designers of DDs poorly understand
this problem. Few efforts to design appropriate mechanisms have
been particularly successful, although there are some cases
of note. Reviewed below in brief are the key institutional frameworks
for participation, accountability, and more effective information
flows between LGs and the public. These mechanisms tend to promote
participation, transparency and accountability together, although
a variety of tradeoffs and complications occur:
- Size of Local Governments
Large LGs are expected to facilitate the provision of
basic community, and in some cases, regional services. Many
countries have learned though that large size is not necessary
to take opportunities of scale as such opportunities can
be developed though contracting and joint production of
services. On the other hand, small local governments have
often led to the creation of several tiers of local authorities,
with serious implications for available fiscal and human
resources.
- Local Assemblies
Local assemblies are expected to provide oversight over
the executive. In many instances, local executives exercise
much more power. As with the central legislatures, local
assemblies are often incapacitated to make contributions
to the policy process or to hold the executive accountable.
A variety of instruments have been developed to tackle this
problem in different countries and will be reviewed.
- Local Elections
Compared to national elections, local elections have not
been seriously studied. Some countries have not even allowed
multiparty elections at the local level even though they
permitted political competition at national levels. Elections
constitute a major form of participation, as indicated by
voter turn-out figures. They are also critical for accountability,
whether or not representatives are responsive or accountable
to their electors. Finally, they provide information to
the electors. The constituency mode is said to assist in
terms of equitable geographical representation of the local
electorate whereas proportional representation helps reflect
associational forms more effectively. Key arguments on the
extent to which local elections reflect majority versus
minority interests are reviewed. This will also lead us
to an examination of the arguments for elective and selective
councils, for party-based versus non-party based LGs and
for constituency based versus proportional representational
forms at the local level.
- Political and Administrative Structure of the Councils
There are different council management systems. The most
important forms are the weak mayor and the strong mayor
systems. In the latter, the whole community elects the mayor
or chair of the councils into the council who serves in
an executive rather than ceremonial capacity. Other accountability/transparency
mechanisms reviewed include the following: forms of voice
options- civil society, media, public meetings, formal grievance
procedures, opinion surveys; budget tracking experiments.
Exit options have been used in a few countries (Cote D'Ivoire).
A related issue is that conventional forms of accountability,
which are usually indirect, are being supplemented or replaced
in many countries by direct forms of accountability (e.g.
participatory budgeting, service delivery surveys). The
two Municipal Development Programs one each for southern/eastern
and western/central Africa - have made these major aspects
of their program of assistance to local governments.
In the full paper, the following are also highlighted: Participatory
Budgeting; Social Funds; Integration of Informal and Formal Structures
of Governance; New School Management Approaches; Community Banks;
Media; and Joint Action Planning
Conclusion: The Challenge of Polycentric Local Governance
This final section pulls together the lessons from the cases.
It examines the prospects and constraints of polycentricity
in Africa, given political and economic contexts that have led
to the convergence of Francophone and Anglophone traditions
away from earlier commitment to monocentricity that has brought
so much misery to the continent.
Questions for Discussion
- What legal, cultural, political, economic factors explain
decentralization in the countries with which you are familiar?
- How effective are the following mechanisms in promoting
central-local relations: constitution, revenue transfers from
national to local governments, local government service commissions?
What factors influence effectiveness? Should local elections
be party or non-party based? Why?
- Has DD resolved or aggravated differences between different
parts or groups within the same country?
- What policy/institutional instruments are available to tackle
this problem where it exists?
- Are social funds and joint action organs more effective
in tackling poverty in rural communities when they work with
local governments? How can local governments become effective
institutions in working with these institutions?
- What factors explain the effectiveness of different models
(integrated, separated and unified) of human resource management
in different countries?
- To whom are local governments accountable - the central
government or their own citizens? What are some of the major
benefits and weaknesses of local governments' accountability
to central governments?
- How can greater accountability to the electorate be secured?
In answering this question, consider some of the novel ways
by which the indirect forms of accountability are being supplemented
with direct forms at the local level. What factors explain
the differences in the effectiveness of indirect vs. direct
forms of accountability?
- Local governments are theoretically expected to be more
transparent than central or regional governments. In many
countries they are not. What factors foster opaque local governments
and what has been done successfully to tackle these problems?
- How effective are the following institutional mechanisms
in enhancing the goals of DD in any program of decentralization
with which you are familiar: Decentralization secretariats;
Ministries of Local Government; Planning Agencies; Local Courts;
Traditional Chiefs; Audit Commissions; Community Development
Organizations; Parents Teachers Associations; Social Funds;
Service Delivery Surveys.
Plenary One Discussants
Comments
Mr Roger Shotton, Technical Coordinator, UNCDF
Mr. Mamadou Diouf, Director of Local Government, Senegal
Professor Walter Oyugi, University of Nairobi, Kenya
Specific Comments
Issues of Institutional Architecture and Accountability
On issues of institutional architecture and accountability,
it is generally accepted that accountability mechanisms are
the key to unlocking the service delivery and local governance
potential of local government. Generally there are two types
of accountability issues which require scrutiny, the first is
that of the accountability of local elected officials and the
second, that of local civil servants.
Accountability of local elected officials (Downward Accountability)
This form of accountability is generally agreed to be of the
kind of elected officials towards their citizens, the importance
of which has already been highlighted by several of the speakers.
Related to this form of accountability are various forms of
institutional mechanisms that can strengthen or weaken it. The
first of these is the electoral mechanism, while participation
in elections is important, the electoral mechanism itself can
greatly affect the responsiveness of elected officials to their
constituency and community demands.
- The basis for electing councils for example, as discussed
Professor Olowu, a distinction can be made between a ward
based system where each ward within a local government area
elects a councilor; and a party list or a proportional representation
system. On the face of it, the ward based system would seem
to offer better hope of downward accountability if only because
citizens know better whom to complain to, and may even be
able to establish direct contact with the councilor in question.
Professor Olowu argues that conversely proportional systems
may encourage more associational activities in the area, this
question could be the subject of further discussion as to
the positive and negative attributes of proportional representation
at the local level.
- Mode of Election of the Council Chair or Mayor, which can
greatly affect the accountability of local government. There
are two main options, the first where all citizens in the
area elect the chair or alternatively where there is an indirect
election of the chair by the council. The direct election
option can often result in an over-dominant chair who because
of the strength of his or her power base can overshadow other
councilors and somewhat discourage local democratic debate.
This is a subject of discussion worthy of further debate.
- A third electoral mechanism issue is the extent to which
special provisions are made to ensure representation of groups
who would otherwise have been under represented, typically
women but also transient groups such as unemployed laborers,
pastoralists, or temporary immigrants.
There are three other sets of issues:
- The size and population of a local government area, which
also affect downward accountability potential. There is always
a trade-off between arguments for fiscal and economic viability,
which would favor a larger geographical area, and political
coherency, which would tend towards creating smaller local
government areas. The accountability challenge of larger districts
such as those in Malawi where districts would include 500
000 people; and those of, for example, Mali where the smaller
communes would include only about 15 000 people, where in
the latter case many people know their representatives. This
also raises issues for discussion.
- The clarity and definition of the roles of local government
and elected representatives and the extent to which these
are known by the public, the greater extent to which this
is the case, the greater the pressure for accountability.
- With regard to local planning and consultative procedures,
the extent to which information is made available to people,
budgets and other related information, the greater the ability
to give voice.
There are also a number of systemic issues that may be largely
outside the control of local policy-makers, which can affect
downward accountability. The first is the nature of the party
political system. Conventional wisdom holds that multi-party
politics makes for more downwardly accountable government, the
real issue is the degree of choice and competitiveness between
candidates, a one party system which has competitive and locally
managed pre-selection procedures may be offer more real choice
than a multi-party system where candidates are designated by
Party head-quarters, this is a controversial issue but nonetheless
one which should be taken into account.
A final systemic factor has to do with the strength of local
civil society. Conventional wisdom suggests that where civil
society is stronger there will also be stronger pressure on
local government for accountability. However, the development
of civil society (lobby and pressure groups) will itself be
encouraged where there is a viable local government worth the
time and effort to lobby and pressure. Thus the question sometimes
posed by donors as to whether we should be supporting local
government or local civil society can often be a politically
naive one.
Accountability of Local Civil Servants (Horizontal Accountability)
This area of accountability is often neglected. The term itself
refers to two rather restrictive assumptions:
- It supposes that local civil servants are solely accountable
to local elected representatives rather than to the public.
While this may be realistic for higher level civil servants
it seems to neglect the scope for performance monitoring of
frontline civil servants teachers, health workers and extension
agents by the public itself.
- It assumes that local elected officials and the local executive
branch are at the same institutional level. This is certainly
typical of the District set up in most East and Southern African
countries, but it is not true at all for most francophone
countries where deconcentrated civil servants are situated
at a higher level (the Department, Cercle, or Arrondissement)
than the elected communes, and in no way fall under their
control.
Two additional Myths and Conceptions:
- Elite Capture of Resources
Reference has been made to the view that decentralization
leads to local elite control of resources the suggestion
being that this works against poverty reduction and equity.
While the tendency for elite control is true, political
science suggests that people who are relatively better off
educationally, financially and who have more time and incentives
to participate in political activity control all political
institutions. The real questions to be asked go further,
i.e:
- Are other service delivery institutions such as line
ministries less vulnerable to elite control or access;
- Which type of elite or elite coalition takes control;
and, more importantly:
- What mechanisms can be established for ensuring accountability
and responsiveness of this local elite to local citizens.
- Local Capacity Problems
The allegation of lack of local capacity is frequently
made by both central government and donors as an excuse
for not allocating functions or resources to local government.
While there are very real problems in this respect there
are also misconceptions, for example:
- Often bigger much better resourced local governments
perform no better and often much worse than small, apparently
weaker, local governments, this fact should suggest that
there is more to "capacity" than having trained personnel
in place;
- Local governments do not have to have in house technical
staff associated with every service delivery function.
Technical support can often be brought in from other agencies;
- In many places lack of capacity is not as much of a
problem as inadequate procedures and guidelines; and in
some cases it can also be a question of incentives for
good governance;
- Incentives for good performance may not be clearly stipulated.
Local Governments thus need to be given resources required;
bypassing them is often very problematic.
General Comments
- Time is very important when we deal with decentralization.
A great danger is to forget that decentralization is a long
process, and to believe that we can copy each other's policies,
changing the name of the country and declaring that we have
the same decentralization policy as our neighbors. This is
extremely dangerous;
- Time has a double impact. It is important when we take it
into account to study the context of countries and to know
what we have to consider and what we have to put aside. But
it can become an enemy if we believe that we have enough time,
it takes away the sense of urgency which is necessary;
- The absence of knowledge and control over the context make
us waste a lot of time. It is good to know that the issues
we are confronted to (unemployment, poverty) are not cyclical
but structural;
- We have to take specific situations into account and know
how to adapt decentralization policies to our conclusions;
- We also need to question state justifications for decentralization
policies. The allocation of resources to human needs is central.
It has to be oriented to social and economical development;
- Local taxation methods have to be transferred to the local
level, but in addition there is a need to focus on the disparities
between so called "rich" and "poor regions". This task belongs
to the central state level, and should precede decentralization
of taxation mechanisms;
- When policies implemented from the top are directly imposed
on local populations, decentralization doesn't work. The state,
civil society and development partners have a key role to
play;
- In terms of the relationship between decentralization and
participation, the problem with participation is different
perceptions of participation. Some politicians see participation
as a means of cooptation, for example, and it may be that
political representatives also have vested interests which
make the question of local participation as something reflecting
the will of the people problematic;
- If one looks at the African experience, one finds over the
years, especially since the 1960s, that African countries
have been experimenting with one form of decentralization
after the other. These experiments have ranged from integrated
systems approaches to partnership approaches, without sufficient
analysis of what has gone wrong. These designs have been imported
from outside without much attention as to whether the design
is appropriate to the state in question;
- Both presentations were concerned with mechanisms for mediating
central local-relationships, which is a major question of
policy instrument design. This raises another fundamental
question: what is the principle of the location of functions
between the center and the localities where the understanding
of the localities is sufficiently broad to include a number
of areas of authority. Thus whatever responsibility is assigned
to a level; it is capable of fulfilling that function. There
is however a tendency to want to universalize and generalize
so as to beable to apply the same model everywhere, regardless
of the specific context of states;
- There is an assumption that decentralization is a good thing,
because people below national level wish to share power. This
assumes, that there are those at the center who are willing
to share power with the localities, and the administration
of power sharing is allowing people at local level to make
decisions which have a bearing on them directly.
- Decentralization is not a project or a program. Rather it
refers to the evolution of democratic political processes;
- Decentralization requires trust between various levels of
government, most particularly between central institutions
and those to which they delegate power. Without trust, power
cannot be shared in meaningful ways;
- Donors should streamline their point of entry in states
to prevent different directives at different levels of government;
- Leadership commitment is a prerequisite and is linked to
trust;
- Decentralization should be shown to provide tangible benefits
to those whom it targets, so to as to ensure their commitment
and participation;
- There is a need to examine the genesis of African states
in political and geo-political context, in order to evaluate
the fears which states' face with respect to state-building;
- Illiteracy poses a major problem to ensuring effective participation
in the decentralization process.
Return to top
Plenary Session Two: Fiscal
and Economic Issues
Chair
Hon. Brigadier Ngwilizi, Minister of State, Regional Administration
and Local Government, Tanzania
Presenters
Professor Remy Prud'homme, University of Paris XII, France
Professor Bert Helmsing, Institute of Social Studies, the Netherlands
Discussants
Dr Theo Rauch, Advisor, GTZ, South Africa
Dr Francois Yatta, Regional Advisor, MDP, Benin
Dr Dick Odur, Director, Local Government Finance Commission,
Uganda
Paper Presentations
First Presentation
Fiscal Decentralization
and Intergovernmental Fiscal Relations
Rémy Prud'homme, University Paris XII
Although the focus of the paper is on "fiscal decentralization"
and "intergovernmental fiscal relations", the paper will also
try to take a broader view. Decentralization is about the role
that sub-national governments can and should play in a country's
political and economic life. The important part of the paper,
and the one on which workshop discussions should concentrate,
is the last one, on decentralization strategies. But such discussions
might be facilitated by an initial review and analysis of fiscal
decentralization.
The paper first identifies five main "decentralization
instruments" (the transfers system, for instance), that
can be, and are in fact, combined to produce a decentralized
system. It then identifies four main "decentralization criteria"
(such as economic efficiency), four policy areas that are
impacted, positively or negatively, explicitly or implicitly,
by the decentralization system and its five components. Understanding
that there are many criteria or goals relative by which a decentralization
system must be assessed is important and often forgotten in
many discussions of the "pros" and "cons" of decentralization.
The next step is to examine the "decentralization mechanisms"
, that is how each of our five instruments impacts on each
of our four criteria (the impact of the transfer system upon
economic efficiency, for instance). An understanding of the
most important mechanisms is a prerequisite to the discussion
of polities and strategies.
Only then will we discuss possible "decentralization strategies"
: what could be done with each of the five decentralization
instruments, or to put it differently, what decentralization
system could be designed and implemented in a given country.
Five Decentralization Instruments
To design a system of governance involving two or more levels
of government, social and political decision-makers can play
with several instruments, several policy variables. They are
like a composer writing a symphony for a number of instruments;
the quality of the symphony will depend upon the melody written
for each instrument and also upon the combination of the many
melodic lines. Just as a composer must know each of the instruments
playing in his/her symphony reasonably well, a policy maker
must know the decentralization instruments at his/her disposal.
We identify five main instruments:
1. Allocation of responsibilities
Public goods and services should be provided by public bodies
(who can in turn, in certain cases, contract it out to private
entities). The issue is, to which public bodies? To the central
government? To local governments? To intermediate governments?
The answer to this allocation of responsibilities question is
easy for certain public goods: national defense should be the
responsibility of the central government, and street cleaning
that of local governments. But it is difficult for many, if
not most, public goods. Some broad guidelines can be given.
One can also note that in many cases, responsibilities for
a given service can, and in certain cases should, be shared
between levels of governments. Education is a case in point.
Local governments should be involved, because they are closer
to local conditions and peculiarities. Central governments should
also be involved for political and economic reasons, because
students are future producers and citizens. Intermediate governments,
where they exist, also have a role to play. In such cases, the
problem of allocation of responsibilities between levels of
governments is also one of coordination and cooperation between
levels of governments.
2. Allocation of taxes
Public goods and services must, at least in part, be provided
for free, and financed by taxation. There is a limited number
of taxes: taxes on personal income, on corporate income, on
value-added, on property, on activity, on wages, on the movement
of goods, excise taxes, and so on. The limitless imagination
of Ministries of Finance tends to lengthen the list, but does
not in itself solve the allocation of taxes problem. Which level
of government will be given access to each of these various
taxes? Some taxes may even be allocated simultaneously to several
levels of governments. International experience and theory suggest
that some allocation patterns are better than others. Allocation
of taxes to rural areas is a particularly difficult issue.
Because rural areas are usually very poor, they have very
low tax bases, and they can only impose low tax rates. There
are no good tax bases for rural areas. Taxes on income and wages
are particularly inappropriate; few people have wages, and incomes,
which are low anyway, are not known. Taxes on consumption goods
are not very good either, because much of what is consumed is
generally self-produced or bartered, and very difficult, not
to say impossible, to assess. Taxes on business activities assume
that there are businesses, which is often not the case. Taxes
on goods that are used as inputs for agricultural production
(such as fertilizers) should be discouraged, because they might
lower agricultural output.
Taxes on agricultural goods produced should not be ruled out;
they are proportional, not regressive; they might reduce agricultural
output if they create an price effect (lower price, lower output),
but they might also tend to increase output if there is an income
effect (more output to maintain the same income level). The
best, or least worst, tax base for rural areas is probably the
value of land and property. It is difficult to assess and to
implement land taxes, but land taxes are the least damaging
form of taxes, because they do not discourage output.
3. Transfers system
The standard theory of decentralization suggests that responsibilities
and taxes should be decentralized jointly, and that each level
of government should finance its own expenditures. This is never
achieved, and for good reasons. In all countries, responsibilities
are more decentralized than taxes, and funds are transferred
from the central government to sub-national governments. This
can be done in many ways, and an intergovernmental transfers
system has to be designed. The menu from which one can pick
transfer systems is large and diversified.
4. Central government controls
Decentralization is about transferring power (particularly
but not exclusively in the form of responsibilities and taxes)
from the central government to sub-national governments. This
transfer of power, however, is neither total nor complete. Everywhere,
central governments retain some form of control over what local
governments do, for example, in the areas of expenditures, borrowing,
2. taxation, budgeting, accounting, reporting and auditing.
5. Local government election rules
There is no decentralization without relatively independent
and politically responsive local (or more precisely: sub-national)
governments. The rules that govern the formation and functioning
of local governments are an integral part of decentralization.
They include election rules (how are local councils and mayors
elected; for how long; how many times), but also administration
rules (for example, how are administrators recruited, promoted,
hired and fired; how are budgets prepared and implemented).
They also include the sub-national structures (how many levels
of governments; how large in terms of area and population are
communes, provinces and regions?). All too often, in a given
country, these variables, which have been defined for long,
are taken for granted. They should not. History might make them
difficult to change, and should not be ignored, of course. But
these politico-administrative elements are not intangible, and
should be considered in any decentralization strategy.
Four Decentralization Criteria
Decisions taken about the five instruments just mentioned
will have profound impacts in a number of socio-economic domains.
We identified four such domains, or objectives, that constitute
the criteria by which a decentralization system or reform must
be gauged.
1. Economic efficiency
Economic efficiency has two meanings: allocative efficiency
and productive efficiency. Allocative efficiency refers to the
adequacy of the mix of goods and services provided to the mix
of goods and services wanted by the population, and therefore
to its satisfaction. Allocative efficiency is clearly linked
and usually benefits from decentralization. Productive efficiency
refers to the resource cost of the production of a given public
good. Whether it is increased, or decreased, as a result of
decentralization is a matter of debate.
2. Political efficiency
A decentralization system has important political consequences
and implications. First, increased de-centralization facilitates
and increases local participation in political decision-making.
Even if the purely economic outcome was not modified, this increased
local participation would bring, by itself, social benefits.
People prefer decisions they take to decisions that are imposed
upon them. Second, active local governments are schools of democracy,
for both the electorate at large and for politicians. People
get used to discussing social issues, understanding the nature
of choices offered, realizing the need for trade-offs, and becoming
better-informed citizens in general. Local governments are also
effective training grounds for politicians. Finally, stronger
sub-national governments are a useful counter force to the central
government. They provide an additional protection against the
temptation of authoritarianism at the center. All other things
being equal, it is more difficult to become a dictator in a
decentralized country than in centralized one. These political
objectives cannot be ignored in a discussion of decentralization.
Indeed, in many cases, such considerations have been the main
motive of decentralization reforms.
3. Macro-economic stability
Macro-economic management, in terms of growth and inflation,
is a major responsibility of government at large, which is also
impacted by the extent and forms of decentralization.
4. Redistribution
It is also generally agreed that income redistribution between
persons and between areas is a major goal of public policy.
The impact of decentralization upon this goal, which is important,
must be taken into consideration.
Decentralization Mechanisms
Each of our five policy instruments has impacts upon each
of our four policy areas. Such impacts must be studied and understood.
Table 1 presents a broad picture of what is to be examined.
The signs in each of the cells are a gross and crude description
of the nature of the impact (with ++ for very positive, + for
positive, 0 for neutral, - for negative, and for very negative).
This description is bound to be very crude, and perhaps even
misleading, for two reasons.
- Country-dependent impacts
One is that these impacts are very much country-dependent.
They are a function of the income level, the administrative
capacity or tradition, the geography, etc. of each country.
What is true in a country with a French tradition is not
necessarily true in a country with a British tradition.
What is true in a small, densely populated, country like
Burundi is not necessarily true in a large, sparsely populated
country like Niger. What is true in a US$ 3,000 per capita
country like South Africa is not necessarily true in a US$
300 per capita country like Madagascar;
- Instrument-dependent impacts
The other aspect is that each of the five instruments
is multi-faceted, qualitative, undefined. There is not just
one type of transfer system, but many different types, and
the problem is precisely to choose the most appropriate
as a function of its impacts upon the four policy objectives
identified. The main virtue of table 1 is to provide a structured
checking list of the questions that should be asked about
a particular actual or proposed) decentralization system,
and not in the tentative answers that are given. These tentative
answers, however, which are given for an "average" decentralization
system (with an average set of election rules, an average
allocation of responsibilities, etc.) suggest that decentralization
is rather good for political efficiency, probably good for
economic efficiency, and rather bad for macro-economic stability
and redistribution. Table 1, therefore should be completed
for each country and each set of policy instrument chosen.
Such a task might be facilitated by an understanding of
some of the mechanisms that are at work;
Table 1 - Analytical Framework for Impact Mechanisms
| |
Economic
efficiency |
Macro-economic
stability |
Redistribution |
Policital
efficiency |
| Allocation
of responsibilities |
+ |
- |
- |
++ |
| Allocation
of taxes |
- |
-- |
-- |
++ |
| Transfers
system |
+ |
+ |
++ |
-- |
| Central
government controls |
+ |
+ |
0 |
0 |
| Election
rules |
+ |
-- |
- |
++ |
| All |
+ |
- |
- |
++ |
- Disparity enhancing mechanisms
One such mechanism is the adverse impact of decentralization
on interpersonal and interregional disparities. All other
things being equal, a more decentralized system will tend
to increase disparities. The initially poorer local jurisdictions
will have lower tax bases (whatever the taxes allocated
to them), and therefore lower resources, which means lower
levels of public services, that will tend to evict people
and activities, making these areas even poorer. Such a tendency
does not condemn decentralization, of course. It might be
accepted as a price to be paid, and/or corrected, at least
in part, by a transfer system that gi
