Local and Community-Driven Development
Historical Roots and the Evolution of World Bank Programs
Hans P. Binswanger
Steen Lau Jorgensen
Jacomina de Regt
March 11, 2006
1
Table of Contents
Introduction ......................................................................................................................... 3
Box 1: Linking Community Empowerment, Decentralized Governance, and Sector
Approaches to Local Development ............................................................................. 5
Part 1: A sampling of historical experiences ...................................................................... 6
1.1 Sector-specific approaches........................................................................................ 6
Box 2: From Specialized Agricultural Credit Institutions to Micro-Finance ............. 8
1.2 Community Development in India............................................................................ 9
Box 3: The Comilla Model of Rural Development .................................................. 10
1.3 Technology-led production programs ..................................................................... 10
1.3 Technology-led production programs ..................................................................... 11
1.4 Programs for special areas and special target groups ............................................. 12
Box 4: Slow progress in Decentralization ................................................................ 12
1.5 Integrated Rural Development and Area Development Programs ......................... 13
Box 5: Participatory appraisal and planning ............................................................. 14
1.6 Non-governmental approaches ............................................................................... 15
1.6 Key Lessons of History........................................................................................... 16
Part 2: Local and Community-Driven Development in the World Bank since 1990 ....... 17
2.1. Towards the synthesis ............................................................................................ 17
2.2 How well are the vision and the tools being implemented? ................................... 20
Box 7: Table ES1 of the OED report here ................................................................ 21
2. 3 How did the separate approaches evolve ............................................................... 22
2.3.1 Participation and de-concentration in sectoral programs ................................. 22
Water supply ......................................................................................................... 23
Urban development ............................................................................................... 23
Natural resource management............................................................................... 23
Nutrition ................................................................................................................ 23
2.3.2 Breaking out of the sectors: Social Funds and AGETIPs ................................ 23
Social Funds .......................................................................................................... 24
AGETIPS .............................................................................................................. 24
2.3.3 Fostering Participation: The Sourcebook......................................................... 24
2.3.4 Decentralization approaches to local development.......................................... 24
2.3.4 Community-driven Development in a Decentralization Context .................... 26
Mexico .................................................................................................................. 26
Brazil ..................................................................................................................... 29
Indonesia ............................................................................................................... 31
Subsaharan Africa ................................................................................................. 31
2.3.5 The development of community-based disbursement and procurement
mechanisms ............................................................................................................... 31
2.4 The future agenda for LCDD .................................................................................. 32
From Messy Boutiques to Scaled up LCCD ............................................................. 32
Applying the guiding principles................................................................................ 33
Scaling up.................................................................................................................. 33
References ..................................................................................................................... 33
2
Introduction
The idea of poverty-reducing rural development, and of involving communities in their
own development, has deep historic roots. Mahathma Ghandi considered rural uplifting
and reconstruction an essential feature of his program. He advocated communal harmony,
economic equity, social equality, abstinence from alcohol and narcotics, promotion of
'khadi' (hand spun and hand woven cloth) and village industries, sanitation, health care,
education and empowerment of women.
In the last sixty years developing countries and their development partners have struggled
to implement this vision. In particular many ways have evolved in which governments
and nongovernmental organizations (NGOs) support local communities. While it is
difficult to do justice to the many different approaches, they can broadly be grouped into
three broad models, that we will use in the later discussion of different development
programs:
1. Under the “Service Delivery model,” government agencies or NGOs consult
communities, but operate as direct service providers via their own staffs. This
model is particularly widespread in the provision of health and education services,
but is sometimes also used where it may be less appropriate. This was the model
which was most often used in agricultural extension services and to deliver
investments and services in many early community development programs and
the area development programs which will be discussed later. In order to produce
the services and infrastructure, a dedicated permanent staff is needed, with the
associated management, incentives, and sustainability problems this implies.
2. Under the “Intermediary model”, government agencies or NGOs work directly
with communities, but take a strong management role in community efforts,
including in project design and choice of technology, the contracting of service
providers, and in the management of funds for the community projects. While this
approach has the potential of gradually building community capacities, in practice
it often tends to limit community empowerment. Because government
organization or NGO take on management functions, they need significant staff
resources and managerial overhead to manage this approach. While NGOs can
use flexible contracting and financial management procedures, government
agencies have to operate under government accountability rules, making the
contracting and disbursement processes unwieldy.
3. Under the “Empowerment model” the government agencies or NGOs operate
primarily as facilitators and trainers to enable communities to identify their own
priorities and carry them out. Communities are heavily involved in the design and
choice of technology for their project. For small projects they usually manage the
project money and directly contract for goods and services to implement them.
3
The successful Aga Khan Foundation programs in Pakistan and India, and the
Onchocerciasis eradication program in Africa utilize this approach. Communitydriven development, as defined today by the World Bank, also uses the
empowerment model.
Community-development does not exist in a vacuum, but instead is a component of
broader local and national development. An International Conference on Local
Development held in Washington in June 2004 was the culmination of a long process by
which various approaches to local and community development and service delivery
were synthesized into a coherent framework. Box 1 provides excerpts from the Local
Development Framework Paper written for this conference (World Bank 2005b).
4
Box 1: Linking Community Empowerment, Decentralized Governance, and
Sector Approaches to Local Development
Three alternative approaches to local development—decentralized sector approaches, local
government approaches, and community support approaches—emphasize many of the same
principles: empowerment of the poor and other marginalized groups, responsiveness to
beneficiary demand, autonomy of local institutions, associated with greater downward
accountability, and enhancement of local capacities.
Although the three approaches all aim to provide public facilities and services at the local
level, they organize their efforts differently. While sector approaches organize according to
the functions to be performed or the services to be provided, local government approaches
organize based upon the territorial jurisdiction under a legally autonomous authority. The
basis of community support approaches is the social unit through which people organize,
either traditionally or voluntarily, to make and act upon collective decisions. As a result of
these fundamental differences, each approach has generated a distinct body of theory and
practice relevant to supporting local development.
To complement the services provided by sectoral agencies and local governments, programs
using highly decentralized, participatory and demand-based methods such as community
driven development (CDD) have demonstrated considerable success in getting resources to
their intended beneficiaries and in achieving rapid impacts.
Community
Support
Approaches
Local
Government
Approaches
Linked
Approaches
Decentralized
Sectoral
Approaches
A decentralized and integrated approach involves organizing interventions around local
territorial units such as districts, municipalities, or communes. Community-based
organizations (CBOs), local governments, and deconcentrated sectoral agencies, as well as
private organizations such as NGOs and firms, should be linked more coherently in order to
support improved empowerment, governance, service provision, and private sector growth. A
spatially framed approach, which links such local organizations through their respective roles
and relationships at local government and community levels [emphasis added], promises to
improve coordination, synergy, efficiency, and responsiveness in local development
processes.
Source: World Bank, Local Development Framework Paper, 2005
5
To achieve an integrated local development outcome, the new consensus makes
community-driven development and integral part of local development, and involves the
support of sector agencies and programs. What was emerging in this process could no
longer be captured by the term “Community-Driven Development” but instead had
become a framework for Local Development. Therefore we will are now consistently
using the term “Local and Community-Driven Development (LCCD),” a terminology
which clearly states who is supposed to be in empowered and be in charge.
In part 1 the paper will provide a sample of historical examples of agricultural and rural
development which have been tried and scaled up until about the 1990s. The concepts
and vocabulary just discussed will be used to make this section more understandable. As
we shall see, the development of the recent consensus was far from linear, and instead
characterized by brilliant blueprints and pilots, by significant successes and large scale
failures, and by protracted intellectual battles among the advocates of alternative
approaches. Most strikingly we will see that the key ideas of combining empowerment,
decentralization, and sector-specific expertise have long historical roots. This section
closes with a set of conclusions and lessons.
Part 2 of paper then deals with the Evolution of Local and Community-Driven
Development over the past decade and a half, the slow emergence of the consensus just
discussed, and the progress in implementing the new consensus. Of course the World
Bank effort built on the global experience and its failures and successes.
Part 1: A sampling of historical experiences
1.1 Sector-specific approaches
Historically, systematic government support to agricultural development started with
sector-specific approaches and programs which were usually focused on increasing
agricultural production rather than poverty reduction, and tended to take advantage of the
best agricultural opportunities. Today’s enormous canal irrigation systems in South Asia,
China, Egypt, Mexico, Brazil are the legacy of sector-specific irrigation bureaucracies.
From 1965 to 1986 irrigation projects were about 25 percent of the total agricultural and
rural development lending of the World Bank (World Bank, 1987). Other sector-specific
approaches involve agricultural research and extension, roads and water supply, health
and education, forestry, land administration, and targeted agricultural credit (Box 2).
They all used the service delivery model of community involvement, and almost never
involve local governments.
As discussed in the World Development Report of 2003a on Service Delivery, many of
the sector-specific central bureaucracies have failed their users, and in particular the poor.
Many of these agencies were and are still far too centralized and, not accountable to the
users. Some are engaged in rent seeking, if not in outright corruption. They spend a
disproportionate share of their resources on staff and offices in the capitals, and have
rarely been capable of delivering services to the rural poor. In many countries efforts are
underway to make these agencies more accountable to users, de-concentrate their staffs
6
and services, or devolve their part or all of their direct service delivery functions to local
governments or communities, such as irrigation associations.
7
Box 2: From Specialized Agricultural Credit Institutions to MicroFinance
Specialized agricultural credit institutions were one of the most important
sector specific programs set up all over the World. India, for example,
developed a comprehensive system of cooperative land development banks for
investment credit and of agricultural credit societies for seasonal input credits.
These systems built cooperative savings and loan associations, which they then
federated into the larger systems. From 1965 to 1973 about a quarter of World
Bank agricultural lending was for agricultural credit, with the largest financial
support to India, Pakistan, Brazil, Mexico, and Morocco. These credit
institutions were used by the other programs which are discussed in this
chapter, including integrated rural development programs. Almost no selfsufficient institutions emerged out of this approach, and most systems ended
up with a large number of dysfunctional and bankrupt primary credit societies.
Moreover, most of the benefits from credit subsidies provided by these
systems were captured by the rural elites, the better-off and well connected
farmers. These failures led the World Bank to stop supporting specialized
agricultural credit institutions. (World Bank, 1987).
In the 1990s the focus shifted to micro-finance institutions, which emphasized
savings as much as credit, and which had poverty reduction as a main
objective. Spectacular example of scaling up are the micro-finance institutions
in Bangladesh, such as the Grameen Bank, or the Bangladesh Rural
Advancement Committee (BRAC). The high population density helped reduce
transactions costs, and the high agricultural intensity, often with three crops
during the year, increased economic opportunities year round. In other areas of
the World where population densities are lower and agriculture usually
depends on rainfall in a single season, the micro-finance institutions have not
been able to scale up to the same extent. Further scaling up of micro-finance is
one of the major objectives of declaring 2005 the Year of Micro-Finance.
8
1.2 Community Development in India
In 1948, pilot community development projects were set up in India under the leadership
of S. K. Dev, the first minister of community development. The pilots were clearly
intended to empower local communities to undertake their own development. The pilots
were scaled up into a national community development program which by the end of the
First Five Year Plan (1952-57) included 1114 blocks covering 163,000 villages, and by
the 1960s the entire country was covered. (Hedge, 2000). The program aimed at
upliftment of the rural poor, covered agriculture, animal husbandry, roads, health,
education, housing, employment, social and cultural activities. Efforts were made to
improve literacy, health, sanitation, housing, transport and communication.
A de-concentrated service delivery model was used to implement and scale up the
approach. An extension organization, headed by a Block Development Officer (BDO)
was established at each block or sub-district, with a team of subject specialists and village
level workers (VLW). The village level worker brought knowledge to the villagers. Each
VLW covered 5-10 villages and assisted with the implementation of various development
programs launched by different departments.
From 1950s till mid 1970s, there were few achievements. Contrary to the intentions of
the pilot projects, community empowerment and democratic decentralization were
minimal. During the initial phase of community development, government officials
prepared the plans. These plans were to be implemented by the sector-specific agencies,
which, as discussed above, suffered from many weaknesses. Communities at best were
consulted, rather than in charge of resources and implementation. In addition the
resources allotted for community development during 1952-67 were so low that it worked
out to hardly to 8 rupees (between one and two dollars at the then prevailing exchange
rates) per capita over this period (ibid). In Box 3 the Comilla Model of Rural
development is discussed, whose early intentions were also subverted in the scaling up
process
9
Box 3: The Comilla Model of Rural Development
The Comilla model emerged in the mid 1960s, engineered by the Bangladesh Academy for
Rural Development (BARD), which was led by Akhter Hameed Khan. The assumptions
behind the Comilla Model are that (i) the problems of rural development should be
approached from the villagers' point of view, because they have the best understanding of
their problems; (ii) they are capable of bringing about changes in their conditions if they have
been provided with the means for development; (iii) agricultural development should be made
an essential step in initiating a broader rural development process; (iv) training, research and
demonstration are essential in promoting rural development.
The most important element of the Comilla approach was the creation of an institutional base
in rural society, and then integrate around it certain basic development programs. The
institutional base consisted first of a two-tiered system of agricultural cooperative to
encourage farmers to generate capital by thrift deposits, help them to get supervised credit,
and to encourage farmers to adopt the technological innovations jointly. Other salient features
of the institutional framework include (i) involvement of both public and private sectors in the
process of rural development; (ii) development of a cadre of institutional leaders in every
village, such as the manager, model farmers, women organizers, youth leaders, village
accountants; (iii) priority on decentralized and coordinated rural administration with due
coordination between officials of various government departments and the representatives of
people's organisations. The Thana Training and Development Centre (TTDC) was designed to
be the locus of the coordination.
The program placed heavy emphasis on economic and technological components of
development. The Thana Irrigation Programme was to provide irrigation facilities to farmers
through their participation in planning and implementation of irrigation schemes. The Rural
Works Programme (RWP), which was the function of the local government institution.
In 1972, the government put in place the Integrated Rural Development Programme (IRDP) to
replicate and expand the Comilla Model in other parts of the country. Later the programme
was transformed into an institution called Bangladesh Rural Development Board (BRDB).
The BRDB eventually became the largest government organisation involved in production
oriented schemes, expansion of the two-tier cooperatives, a large number of target group
oriented projects such as the rural women projects and projects for the rural poor.
By never transferring financial resources and implementation responsibility to communities
the scaled up program drifted back towards the service delivery approach. While the
government introduced elected local governments, it never gave them full autonomy, and also
starved them of funds. The BRDB programs therefore lost many of the decentralized and
empowerment features of the Comilla model.
Source: (Banglapedia 2005b)
10
1.3 Technology-led production programs
As a consequence of the failure of the Community Development program, and the
growing dependency of India on imported food grains, priorities in the 1960s shifted to
agricultural production and a distinct technological orientation was given to the
programs. The Intensive Agricultural District Programme (IADP) supported by the Ford
Foundation, concentrate money, expert staff and agricultural inputs in a few wellendowed agricultural districts (Staples 1992). Production was the goal: equity was not a
major consideration, which caused consternation among those Indians who believed this
violated the tenets of Indian social democracy. The IADP program produced limited and
mixed results. (ibid)
One of the problems of the IADP programs was the lack of technology which provided
sufficient returns to farmers. This constraint was overcome in the mid 1960s with the
advent of the “Green Revolution” varieties of wheat and rice, which emerged from
international centers supported by the Rockefeller and Ford Foundations. The new
varieties spread rapidly in many areas of the developing world. The new short, stiffstrawed varieties raised the returns to improved seeds, fertilizer and irrigation, and
spawned a boom in private irrigation investment based on wells. The benefits of the new
varieties not only helped the incomes of the farmers who grew them. Landless laborers
and small farmers benefited from reduced food grain prices, and from the increases in
labor demand associated with the higher production levels, which reflected themselves in
higher rural wages. These successes led to the creation of the Consultative Group of
International Agricultural Research to support research in all major tropical food crops
via a string of international centers.
It soon became apparent, however, that only nations which invested into their own
adaptive research capacities could take full advantage of the technologies becoming
available internationally, by developing varieties which were fully adapted to their local
conditions. At the national level the Green Revolution therefore led to increased
investment in national research system all over the world, including the creation of
EMBRAPPA in Brazil, often supported via donor funding, especially from World Bank
loans.
While the Green Revolution was the key event which eventually enabled Mexico, China,
India, and most other Asian countries to become self-sufficient in food grains, it became
quickly apparent that the new technology was initially only adapted to agricultural
systems characterized by good irrigation and water control, and was unable to spread to
areas where irrigation was either not available or did not afford good water control, and
for dryland crops and areas and crops. It took a long time for varieties to emerge which
provided a clear advantage in terms of yields and risks in areas of poor water control and
in dryland areas. And up to today, the responsiveness of many dryland crops to fertilizers
remains insufficient, or too risky, and therefore does not entice farmers to invest in
fertilizers, especially where fertilizer distribution systems are poorly developed and
fertilizers are still expensive, as in many parts of Africa.
11
1.4 Programs for special areas and special target groups
As a consequence of the inability of the Green Revolution to solve the poverty issues
outside of the most agriculturally suitable areas, the need for programs with a special
poverty focus reemerged. In the 1970s, India’s central government introduced
independent administrative hierarchies to carry out special programs to reach the rural
poor, and poor areas, bypassing the Panchayati Raj system (Box 4). The targeting
mechanisms in these programs were intended to avoid elite capture of the benefit. They
included the Small Farmers' Development Agency, Tribal Development Agency,
Marginal, Small Farmers and Agricultural Labourers Development Agency. Area
development programs for Command Area Development, Drought Prone Areas, and Hill
Areas were financed and operated directly by the Central Government (Hedge 2005).
Most of these programs used targeted and subsidized agricultural credit as one of their
main instruments to achieve their objectives (see Box 2). Implementation experience was
mixed. Similar programs emerged all over the world.
Box 4: Slow progress in Decentralization
As early as 1957 the Indian government appointed the Balwantrai Mehta Committee
to evaluate the Community Development Program. It found that the program failed to
evoke popular initiatives. The Committee recommended the formation of a three-tiersystem of rural local Government, to be called 'Panchayati Raj' (Rule by Local
Councils), with counsels at the village, the block and the district level. The aim was to
decentralize decision making closer to the people, encourage their participation, and
place the bureaucracy under the local people's control. The Panchayati Raj was
initially not able to fulfill these expectations. Few financial resources were devolved
to the system, which set it up for failure. By the end of the middle of the 1960s
priorities and programs shifted back to centralized modes, and especially to food
production. Almost all the development programs were implemented by the
responsible government departments, and there was little or no scope for participation
by the villagers. Not surprisingly this did not work well and created a dependency
syndrome.
By the late 1980s the limitations of the central agency model to empower and involve
the poor communities became apparent. In India, therefore, movements to revitalize
the Panchayaty Raj system started, culminating in the 73d constitutional amendment
of 1992. These empowered the PR institutions to shoulder the responsibility of
development and decentralized planning. Under this constitutional amendment, 29
items of development were transferred to PRIs, and States were required to allocate
fiscal resources to the system. The central government too started to devolve
implementation for its Centrally Sponsored Schemes to the PR institutions. Presently
over 3.3 million elected leaders have assumed positions in the Panchayati Raj
administration covering 227,678 village Panchayats, 5906 Panchayat Samitis and 474
Zilla Parishads. The reservation of 33 percent of the elected posts for women and
Scheduled casts and tribes has resulted in more active participation of these groups.
Implementation of the constitutional amendment and devolution of fiscal resources to
them still varies widely among Indian States.
Source, Hedge 2000
12
1.5 Integrated Rural Development and Area Development
Programs
By the early 1970s, the initial successes of the Comilla Model, and the failure of the
Green Revolution to reduce poverty in all but the most agriculturally suitable areas,
provided inspiration for the creation of many Integrated Rural Development Programs, or
area Development Programs. Following the famous “Nairobi Speech” of Mr. McNamara
in 1973 the World Bank sharply increased its lending to agricultural and rural
development. The new strategy advocated a continued involvement in sector-specific
programs such as irrigation or agricultural credit, but also introduced a new focus on
integrated rural development via area development project, which strongly emphasized
smallholder agricultural productivity. The ADPs were rural investment programs, often to
serve the rural poor in previously neglected and degraded rural areas. They included
agriculture and other sectors, such as rural infrastructure, health, education, water supply,
and small scale rural industry. Up to 1992 the World Bank assisted nearly 300 such
projects, 45 percent of which were in Africa.
The Rural Development Policy of 1975 emphasized that rural development should be
participatory, decentralized, embedded in a favorable agricultural policy regime, and
based on good available technology. In Mexico, the three PIDER projects implemented
from 1975 to 1988 were considered to be at the cutting edge of a ‘social engineering’
approach to participation (Box 5).
However most of the area development projects financed by the World Bank did not
follow the Bank’s own policy guidance. The participatory planning process followed in
Zacatecas in box 5 was a rare exception. Instead most projects were prepared in a hurry
by teams of agricultural professionals, with little involvement of the beneficiaries. Project
preparation teams, not the beneficiaries and those to be involved in implementation,
prepared the development plans for the areas. Implementation of the different program
components was entrusted to the central or regional government agencies in charge, such
as the extension service, the irrigation department, or the ministry of education. These
inevitably used the service delivery approach. Even in the PIDER projects in Mexico
decentralization did not go beyond the level of the state. This was in clear contradiction
to the policy which recommended a decentralized approach and strong local institutional
development. The central agencies often had priorities which were not necessarily those
of the project designers, or the affected populations. In the PIDER project in Zacatecas,
for example, what was delivered at the community level by each agency was rarely in
line with the priorities of the community identified at the elaborate planning stage.
Implementation by central agencies gave rise to a major coordination problem, which the
World Bank tried to solve by strengthening of the project units, often staffed by
expatriate advisors.
13
Box 5: Participatory appraisal and planning
During the preparation of the second phase of the PIDER project, the Zacatecas State
Development Plan became the first instance when in 1979 the plan of an entire state was
prepared through a participatory methodology. The programming exercise included
diagnostic studies in all communities with populations of between 250 and 3,000; in
addition, communities of more than 3,000 inhabitants were contacted through their
authorities. The survey work itself had two foci--locality studies and sectoral studies. For
the former, sixty field teams using a total of 120 technical experts carried out the diagnostic
work; in the sectoral studies, eight groups including over 100 technicians were responsible
for the analysis. About 80% of the total population of Zacatecas state was contacted by the
field teams. A total of 4,029 investment proposals were received as a result of direct
consultations with communities, and an additional 2,209 projects were proposed by
government departments. After analysis of the survey data, the planners proceeded with the
definition of priority objectives, strategies, specific investments for various communities,
and sector plans. [Cernea, Michael “A Social Methodology for Community Participation in
Local Investments: The Experience of Mexico’s PIDER program”, 1983].
The lighter Rapid Rural Apprasal (RRA) techniques also developed in the 1970s and
1980s. RRA developed its own assessment tools which enabled the outside planners
to gain a better understanding of the development realities of communities than
would be possible by expert visits, and in less time than large scale survey research,
or the process used in Zacatecas. The RRA techniques were soon transformed into
Participatory Rural Appraisal (PRA), which transfers the role of analyzing and
planning to the community itself. It is a structured process to help communities
understand their constraints and opportunities, and develop their own sets of
development priorities. Facilitators help communities to develop, present and
analyze information. Techniques involve (i) diagrams, maps or quantification are
created and presented physically by rural people in a manner they readily
understand. (ii) walking transects across the village to gain a shared understanding
of the environment, (iii) household listings and wealth ranking, (iv) reporting and
analzing the findings in different discussion groups, such as men, women and youth,
(v) ranking and scoring of constraints, options, opportunities and priorities.
Visualization of the results for all to see and understand is a major element of these
techniques. Source: The World Bank’s Participation Sourcebook (2003b).
Participatory appraisal and planning techniques techniques have become an essential
tool for systematic planning and priority setting in community development
programs all over the world, in both rural and urban settings. The results of
participatory rural assessments are often used as an input into participatory planning
at the local level. Participatory appraisal approaches have also been applied widely
to participatory planning at the level of local governments. For example they are
widely used in the pilot local development fund programs initiated by the United
Nations Capital Development Fund in over 15 developing countries (UNCDF,
2005).
14
OED reviewed these programs in 1987 and 1990, and again summarized the lessons in
1993. Performance of the ARD portfolio was significantly worse than that of other
agricultural projects financed by the World Bank, or of projects in other sectors, and was
especially poor in Africa. Overall half of area development programs failed, and in Africa
it was two thirds. Projects were more successful where government commitment was
strong and where the agricultural policy environment was better. Project benefits were
rarely sustainable, and projects attained little institutional development impact, especially
where project management units were used, and even more so where they were staffed by
expatriate advisors. Central coordination of the sector agencies never worked. Locally
proven technologies were often not available, and project-specific technology
development components set up to remedy the situation usually failed. Monitoring and
evaluation was often poor or non-existent. All in all, the projects did not follow the
institutional development approach of the Comilla model, or the core strategic elements
of the World Bank’s own rural development strategy.
In the early 1990s the area development project approach was abandoned by the World
Bank, along with lending for large scale irrigation construction and rural credit
institutions. As a consequence, the rural agricultural and rural development portfolio of
the World Bank shrank significantly, and the remaining programs focused on agricultural
policy reforms, agricultural research and extension, and irrigation rehabilitation. Natural
resource management projects became more important focusing on forestry and
watershed management.
1.6 Non-governmental approaches
Under the pressure from the failure of many state led efforts, governments in South Asia
and elsewhere started during the 1970s to recognize the role of voluntary organizations or
NGOs in supplementing government efforts in rural development. NGOs typically
support the initiatives of local communities. They cover a wide range of rural
development activities including relief and rehabilitation, family planning, mother and
childcare, health and sanitation, income and employment generation, etc . In India, the
Ministry of Agriculture created the 'Freedom from Hunger Campaign' to support
voluntary organizations involved in rural development, which eventually became the
Council for Advancement of People's Action and Rural Technology (CAPART). Success
of these initiatives have now encouraged many state governments to launch schemes to
promote people's participation and several centrally sponsored schemes have stipulated
the development of community based organizations to plan and implement the program
(Hedge, 2000). In many cases, the NGOs act as contractors for government financed
programs. In many countries contracting NGOs has led to the proliferation of
intermediary NGOs servicing pre-formulated schemes, and to an explosion of self help
groups, with little connection to decentralized local or village government. And
opportunities to set up intermediary NGOs have sometimes drawn valuable staff
resources from government services.
15
1.6 Key Lessons of History
The sample of historical experiences brings out a number of major features:
1. All reviewed large scale programs for community development and poverty
reducing rural development started out with strong ideals of community
participation and empowerment, coordination of development efforts at the local
level, flexible and responsive local planning and implementation mechanisms,
and decentralization of decision making and financial responsibilities. Many
also viewed all or parts of the institutional, social, and agricultural policy
environment for rural development as hostile. The sad lesson from this review is
that most of the programs rushed into scaling up long before the institutional,
social and policy environment could have been reformed. As a consequence
many efforts ended up as programs of centralized bureaucracies, which stifled
participation and empowerment, were unable to coordinate implementation of
various sectors effectively, and choked in their own bureaucratic procedures.
2. The large scale programs paid little attention to reforming the institutional
environment, and devolution of functions and financial resources to local
governments or communities, that would have been in a better position to help
coordinate and supervise the development actors on the ground. While India
created elected local governments as early as the late 1950s, it only started
supporting them in earnest in the 1990s. Similar experiences of creating elected
governments, but then not devolving authority and fiscal resources to them
occurred in Ghana, Tanzania, Mexico and other developing countries.
3. When the central sector agencies failed to implement the programs on the
ground, many governments turned to NGOs to help with implementation.
Particularly striking is that most NGOs used the “Intermediary model” and the
“Service Delivery model”, rather than the “Empowerment Model” for supporting
communities. Therefore they often substituted their own staff for the latent or
actual management and implementation capacities of communities and service
users, and failed to seize an important opportunity to reduce overhead costs and
to overcome their own internal bureaucratic hurdles.
4. The early thinkers and leaders fully recognized the dangers of elite capture in
their environments, and tried to build programs which would mitigate these
dangers. Many governments created special targeted programs for the weaker
groups and poorer areas which were managed from the center, or started to
involve nongovernmental organizations in the program administration. These
programs had very mixed results in reaching the poor. It was only in the 1990s
that communities and local governments were starting to be used to target their
own poor members.
16
Part 2: Local and Community-Driven Development in the
World Bank since 1990
During the 1980s the World Bank changed its main focus from targeted poverty
reduction to the management of the debt crisis, via macro-economic and sector policy
reforms. However, by the late 19802 it became clear that the expected positive results of
the adjustment programs for growth and poverty reduction were coming much slower
than hoped for, and in many cases failed to appear altogether. At the same time the
stabilization and structural reform programs had imposed significant losses on poor
groups and social safety nets were urgently required. Fortunately, the World
Development report on Poverty of 1990 moved poverty reduction back to center stage in
the World Bank priorities. The strategy rested on a dual approach of accelerating growth,
complemented by targeted programs for those bypassed by growth.
But, as the historical examples of Part 1 show, the World Bank was in a poor position to
respond to targeted poverty reduction. Central sector agencies had a poor record in
delivering services to the poor, and were often costly and corrupt. Declining international
cereals prices undermined the rate of returns to large scale irrigation programs, and they
came under increasing attacks from environmental and social groups. The large scale
agricultural credit programs had created costly institutions, and had failed to reach the
poor (Box 2). Technology-led approaches proved of limited applicability in resourcepoor areas. The area development programs had proven a failure.
It was therefore not obvious how to deliver the targeted development and safety net
programs which were called for under the 1990 poverty reduction strategy. In this
environment, staff of different sectors experimented on their own, and a number of
parallel approaches emerged which often competed with each other. All these approaches
selectively built on the international experienced reviewed in Part 1. But they did not
systematically use the international experience to develop a unified framework.
Therefore, as discussed in the introduction, it was only recently that the various different
approaches have been brought together into a unified framework of local and communitydriven development (World Bank, 2005b).
2.1. Towards the synthesis
In the World Bank, the process which led to this synthesis started in 1999 (?) with the
creation of a cross-sectoral Working Group on Community-Driven Development (CDD
Group) by President James Wolfensohn. 1 In his many country visits he had seen first
Mr. Wolfensohn’s decision to call for the formation of this working group resulted from a meeting
between him and the following staff from the Africa Region, Daniel Benor, Senior Advisor, Callisto
Madavo, Vice President, Jean-Louis Sarbib, Vice President, and Hans Binswanger, ESSD sector director.
The group presented Mr. Wolfensohn with an initial version of the Africa CDD vision discussed below
(World Bank, 1990), and proposed that it become a major operational thrust in the Africa Region.
1
17
hand what the different community development approaches which were being pursued
could achieve. But he was also bewildered by the many approaches to community
empowerment which were being used. The terms of reference of this group, anchored in
the Social Development Department, were to ……[Steen to fill out]. For the first time the
CDD Group brought together all practitioners of community empowerment and
decentralization approaches to jointly review the many programs and approaches applied
in the World Bank. The CDD group became the instrument for consensus building and
integration of approaches, as well as a key driver in moving the agenda forwards. On its
comprehensive website it currently celebrates its five years existence and
accomplishment.
The first full articulation of the vision of an integrated program along the lines discussed
in box 1 was produced by the CDD working group in the Africa Region (World Bank
2000).2 The vision states “The five main dimensions of CDD are: empowering
communities, empowering local governments, re-aligning the center, improving
accountability, and building capacity”. (ibid, p. 9). Realigning the center was understood
as a reform effort of the centralized sector agencies so that they would devolve
responsibility for delivery of most services to local governments and community levels.3
One of the first tasks of the CDD working group was to define Community-Driven
Development, and various similar approaches, so that it became possible to define more
clearly what was happening in the different Regions of the Bank. The definitions used for
tracking the World Bank’s support to CDD are given in Figure 1. CDD clearly is built on
the “empowerment model” of supporting communities discussed earlier, as CDD is now
defined as programs in which the communities had control over their own projects and
resources. Consultation and participation in maintenance of community projects was no
longer sufficient to classify as Community-Driven Development, but is defined as
……….. Another feature which comes out clearly in the classifications is that efforts to
reform the institutional environment, including decentralization and capacity building for
local governments, were also tracked and given full credit as activities in support of
CDD.
Show the classification here
2
The members of this working group were Hans Binswanger, Jacomina de Regt. Laura Frigenti, Brian
Levy, and Catherine…… The statement was written by Swaminathan Aiyer. In 2001 the Africa Region
followed with a “Sourcebook on Community-Driven Development for Africa” (World Bank 2001e).
Shortly thereafter a similarly comprehensive vision was presented in chapter 9 on CDD on the PRSP
sourcebook, (Dongier et al, 2003).
“Traditionally, central governments in Africa have followed the “blue collar” approach of operating all
services. After decentralization, they will need to shift to a “white collar” approach. Instead of running
services directly, they should focus on facilitating local government activities, setting standards, monitoring
outcomes, providing training to lower levels, and providing rewards and penalties to improve local
government performance.” (ibid, p. 11).
3
18
The classification was used to track progress in CDD, and growth in lending and nonlending support.
Provide data on this growth from 2000 to 2005.
Steen or Jacomina to discuss the trends
A second major task for the World Bank’s CDD group was to develop guidance on how
to implement CDD. The group therefore ten key design principles, that are reproduced in
Box 6. At the same time the CDD Group started developing a number of implementation
tools: for Economic and Social Analysis; Community Mobilization and Capacity
Building; Information and Communications; Monitoring and Evaluation; Targeting and
Selection; Direct Financing and Contracting; Institutional Options, Safeguards, and
Social and Gender Inclusion. It has assembled the existing experience of CDD in post
conflict settings, and in urban development. The World Bank Participation Sourcebook
(2003b) provides a comprehensive overview of methods to enlist the participation of
stakeholders at various levels from the community to local and national level. The Online
Sourcebook on Decentralization provides a comprehensive overview and practical
guidance on how to move forward in this area. CDD has also been adapted and used for
combating HIV/AIDS, managing natural resources, water supply and many more.
19
Box 6: World Bank guidance on key design principles for CDD
1. Establish and enabling environment through relevant institutional and policy
reform.
2. Make investment responsive to informed demand, by providing knowledge
about options and requiring community contributions to investment and
recurrent costs.
3. Build participatory mechanisms for community control and stakeholder
involvement, by providing community groups with knowledge, control and
authority over decisions, and resources..
4. Ensure social and gender inclusion.
5. Invest in capacity building for Community-based Organizations (CBOs)
6. Facilitate community access to information.
7. Develop simple rules and strong incentives, supported by monitoring and
evaluation.
8. Maintain flexibility in design of arrangements.
9. Design for scaling up.
10. Invest in an exit strategy, including permanent institutional and financing
arrangements at an affordable fiscal costs.
2.2 How well are the vision and the tools being implemented?
While perhaps inevitable, the separate development of different approaches, and the slow
process of integration into a coherent framework had a number of adverse consequences.
It led to the application of widely different approaches in projects of different sectors.
They confused country teams, the borrowers and other development partners. The
problem was compounded since other development partners similarly experimented with
new approaches, and rarely coordinated them with World Bank supported projects. Since
the approaches were driven by sector-specialists, these separate projects were not able to
approach the underlying issues of the institutional, social and economic policy
environment. As a consequence they reduced the impact and sustainability of the
projects, as amply demonstrated in the recent OED review of Community-driven
development.
It is therefore not surprising that transforming the vision into sweeping reform and
integration programs within countries has lagged significantly behind the consensus and
intellectual guidance which was being developed. At the level of each country program, it
has proven very difficult to merge the various approaches which were being pursued by
different sectors in the Bank, as well as build the necessary consensus with the central
and sector-specfic government entities, as with programs financed by other donors.
Countries differ vastly in the configuration of approaches which they are using to support
rural development. Some have strengths in their programs at community level, but lag in
decentralization and alignment of the central agencies. Others have made more progress
20
on decentralization, but lag in the area of community empowerment. It is therefore
essential that reform and integration of approaches be designed and pursued
incrementally, starting from where a country actually is.
In a recent review of projects with community participation components which were
approved in by the World Bank between 1989 and 2003, the independent Operations
Evaluation Department distinguished between Community Based Development (CBD)
and Community-Driven Development (CDD).4 The latter support the empowerment of
the poor by giving communities control over subproject resources and decisions, while
CBD gives communities less responsibility and emphasizes collaboration, consultation,
or sharing of information with them on project activities. (World Bank 2005c). Since
1989 the share of projects that include CBD/CDD components grew from 2 percent in
fiscal year 1989 to 25 percent in 2003, and within that portfolio there was a progressive
shift from CBD to CDD. From 1994 to 2003 the outcome ratings of CBD/CDD projects
have been better than for non-CBD/CDD projects, and sustainability ratings have
improved over time. But the interventions have often failed to provide consistent and
long term support and institutional changes needed to achieve long term sustainability.
The projects have done better on quantitative goals such as construction of infrastructure
than on qualitative goals such as capacity enhancement. Those projects that have
supported indigenously matured participatory efforts or when the Bank has provided long
term support to communities beyond the length of a single subproject have done better on
capacity enhancement. The increased access to service delivery infrastructure such as
schools and health centers does not always translate into effective service delivery.
Further the poorest may not always have benefited from these projects. (ibid). OED
provides a very useful overview of CBD/CDD strengths and weaknesses, which is
reproduced in Box 7.
Box 7: Table ES1 of the OED report here
A major failing has been the near absence of rigorous impact evaluation studies and there
is therefore little hard evidence on the impact of the projects on poverty reduction and
community capacity. The projects have enhanced the capacity of government institutions
to implement participatory interventions, but few governments have yet adopted the
approach more widely in their development program. A key recommendation of the OED
report is that CBD/CDD projects still need to be better integrated into an overall
country’s assistance strategy (ibid). Clearly by 2003 the project portfolio fell still short of
implementing the design principles discussed in Box 7. With the exception of a few
OED could not apply the four way classification developed by the CDD group and presented in table ….,
because prior to 2000 the classification did not exist, and could not been applied to projects prior to then.
4
21
projects and programs, the most glaring shortcomings were the area of institutional
reform, full empowerment of communities, monitoring and evaluation, failure to truly
scale up, and development of exit strategies.
Describe the integration and unification efforts which are ongoing in Senegal, Nigeria,
and other countries. Include a reference to the toolkit which Jacomina has produced.
Suggested bottom line: Much has been achieved to spread participatory approaches, but
the rigorous application of the CDD guidelines of the anchor on institutional reform and
development is still not fully implemented. There are still too many sector-specific
boutiques which are not effectively contribution to broader policy and institutional
changes. Of course you can add your own take on the effort.
2. 3 How did the separate approaches evolve
We have already discussed how poorly placed the World Bank was in the late 1980s and
early 1990s to help countries design and implement targeted poverty reduction
interventions. On the other hand the historical section shows that there was available
inside and outside the Bank considerable experience in how to improve participation,
empower communities, make sector agencies more responsive, and build decentralized
government institutions. We will now review how the World Bank progressively
incorporated these approaches into its programs and developed the different strands
which eventually came together in LCCD. Note that the term “Community-Driven
Development” did not exist at the time. It was coined by Deepa Narayan and Hans
Binswanger in 1995 to denote the integration of participatory approaches with
decentralization and direct community empowerment.5
2.3.1 Participation and de-concentration in sectoral programs
World Bank staff in different sectors began to experiment with participatory approaches
in a wide range of sector-specific projects, as well as in area development projects. For
example, from the late 1970s Michael Cernea developed a bottom-up participatory
planning model from the community to the municipal and state level for the PIDER
integrated rural development programs which was widely applied in Mexico. Give more
names and examples.
Below we review a few of the sector-specific efforts
The term “Community Development” was rejected because it was too closely identified with the failed
community development program in India, or the Community development programs in the United States,
which mostly use the “service delivery” or “intermediary” models for working with communities, rather
than the “empowerment” model.
5
22
Water supply
The water supply sector, both outside and inside the World Bank, started early to
introduce a variety of participatory mechanisms in its projects.
Jacomina or Steen to find some early examples
By 1995 Deepa Narayan was able to analyze a sample of 121 completed water supply
projects in forty-nine countries which included various participatory mechanisms. Most
of these projects were completed during the 1980s and involved 18 different bilateral,
multilateral and NGO financing agencies. The study was able to define and measure
participation, overall and at the various project stages. The methods and degree of
participation varied widely among these projects, from representative user committees as
part of service delivery approaches to direct involvement in construction and supervision
of contractors as part of full empowerment approaches. Seventy-nine percent of the
projects were rated low or medium in overall participation and only 21 percent received
high ratings. Multiple regression analysis was used to measure the impact of
participation and other factors. “The results are clear: beneficiary participation
contributed significantly to project effectiveness, even after statistically controlling for
the effects of seventeen other factors…..The proportion of water systems in good
condition, overall economic benefits, percentages of the target population reached, and
environmental benefits rose significantly with participation…..Effective participation did
not result when agencies retained control over implementation details-that is the what,
when, how and where of implementation. At the beneficiarly lever, the two key
characteristics determining participation were commitment before construction, or
demand, and the degree of organization of the beneficiaries [emphasis in the original].”
(Narayan, 1995, pp. 1,2).
Many of the projects analyzed in the above study where small scale, but in the meantime
the water supply sector has done a lot of work to scale up participatory and communitydriven approaches, and in some instances to link it to decentralization efforts to local
governments. Chapter ….in this volume analyzes the early scaling up phase of such a
program in Kerala, India.
Urban development
Kampung Improvement Program in Indonesia
Natural resource management
Jacomina
Nutrition
Steen
2.3.2 Breaking out of the sectors: Social Funds and AGETIPs
While the above examples show that there is significant potential to improve
infrastructure and service delivery in sector-specific programs, the reforms of many
sectors one by one to achieve these improvements was a very slow and hard approach to
23
achieve this, especially since most sector bureaucracies resist such changes. A sector-bysector approach was also not fast enough to deal with adverse employment and welfare
consequences of adjustment programs. Many World Bank staff were therefore looking
for ways to reach communities directly with a broad menu of interventions or assist
municipalities with the implementation of broader infrastructure investment programs
which sector agencies were unable to deliver. Social Funds and AGETIPS (give full
name) emerged from these efforts.
Social Funds
Steen: Boliva, Zambia, Malawi. The shift from intermediary and service delivery
approaches to empowerment approaches. The progressive linking to decentralization and
local government capacity development. The merging of Social Funds and the CDD
agendas.
AGETIPS
Steen
2.3.3 Fostering Participation: The Sourcebook
Jacomina: The first version of the Sourcebook came out in 1995, and was based on a
bank-wide “Participation working group” which originally had been created by Aubrey
Williams, now retired, but still available at awilliams@earthwave.net. He could perhaps write
a paragraph or two on the origins of this effort and the contents of the first source book
2.3.4 Decentralization approaches to local development
Latin America, Tim Campbell, Anwar Shah, and others. And how it was operationalized
primarily in urban development project. How it evolved, including for example the
procect support to the decentralization in Bolivia, Colombia, and elsewhere. The section
should also include a reference to UNCDF pilot programs for Local Development Funds,
which, for example in Mozambique, became the basis for a recent project.
Steen: Can you find someone, such as Tim Cambpell, to write a brief section on this?
Between 1992 to 1996 the World Bank conducted a global study of 14 countries and five
provinces in large countries on “Decentralization, Fiscal Systems, and Rural
Development.”6 The study was designed to complement the work of the participation task
force discussed above, and to focus particularly on the financing of rural development at
the level of local government.
6
The study was headed by Hans Binswanger and the team at various times included Keith McLean,
Graham Kerr, Andrew Parker, Suzanne Piriou-Sall, Johan van Zyl, and Melissa Williams. Outputs are
summarized in McLean et al, 1998, Piriou-Sall ????, and Aiyer et al, eds. 1995,1996.
24
The study examined the level of decentralization in delivering services for six key rural
sectors: primary education, primary healthcare, agricultural extension, rural roads
maintenance, rural water supply, and forestry management. It looked at the powers of
local institutions over delivering crucial services to rural populations.7
The results in the chart below show that, in terms of aggregate sector decentralization,
countries with high agricultural and rural development performance tend to be much
more decentralized than countries with low performance. Jiangxi (China) was the most
decentralized in our sample of nineteen countries/states/provinces. Colombia,
Philippines, and Poland respectively completed the top fifth of countries, and only these
countries scored over 50% in sector decentralization. Imo State in Nigeria received the
lowest score, followed by Cote d’Ivoire, Burkina Faso, and Senegal. Five of the African
countries ranked in the lowest quarter of the sample; Zambia is the only exception.
The high score of Jiangxi reflects views of those who have been working on China over
the past two decades; China is more decentralized than most other countries. This
decentralization is strongly fiscal and administrative and there is considerable devolution
of related powers. In villages and township communities, and there are local non-party
elections to the lowest level elected bodies, Township Peoples Congress and Village
Committees.
7
The study constructed a decentralization index ranging from 0 to 10 based on data collected from World Bank sector
specialists who had worked intensively in the respective countries/states/provinces. It was based on answers to the
following question:
Where is the smallest management unit for rural sector service delivery physically located?
Which level of government responsible for conditions of service of civil servants in the smallest management unit?
How important are elected bodies in (a) service delivery, (b) policy formulation, and (c) funding of each sector?
Which level of government pays the salaries of staff in smallest administrative unit?
What proportion of sector expenditures of smallest administrative unit is derived from the budgets of local
governments?
What proportion of sector expenditures of smallest administrative unit which is derived from user charges, in-kind
contributions, and other beneficiary cost-recovery schemes?
Who determines the budget of the smallest sector management unit?
25
2.3.4 Community-driven Development in a Decentralization Context
The previous section suggests that more decentralized countries/provinces do better than
more centralized ones in agricultural and rural development. However, as discussed in
Boxes two and three, the notion that decentralization must be an important institutional
pillar of rural development goes back to at least the 1950s. Furthermore the Comilla
model explicitly linked decentralized coordination and decision making, local
government and community development. But in scaling up these associations were lost.
It took until the 1990s to until large scale development programs were developed, which
made these links a reality.
Mexico
The Decentralization and Regional Development Projects financed by the World Bank
(DRD I and II) were implemented between 1990 and 2002.8 (Chapter ….in this volume,
on which this section draws, provides much more details on these programs). They
followed on three integrated rural development projects (PIDER I to III) implemented
from 1975 to 1988, and the DRD project was designed to address weaknesses in the
PIDER projects.9 The DRD I project was designed to address each of these. Mexico
8
The leaders of the teams which designed DRD I and II were, respectively, Abel Mateus and Andrea
Silverman. The division chiefs for the two projects were, respectively, Hans Binswanger and Michael
Baxter.
9
The Project Completion Report for PIDER III (1981-88) indicated that, while that project generally met
its objectives, it suffered from several problems. These included, inter alia: (i) low levels of total public
investment in poor regions in spite of the project; (ii) inadequate participation of primary stakeholders; (iii)
26
remains one of the most unequal countries in Latin America, (Gini coefficient of 0.54).
Forty-five percent of the population is still poor, and in rural areas half the population
remains extremely poor. At the time of the introduction of the program, Mexico was
emerging from the debt crisis which hit in the early 1980s, and had led to a massive
increase in poverty, and a sharp reduction of development expenditures. The DRD 1
project was a small component of the broader “Solidaridad” program introduced by
President Salinas to deal with these issues.
The program was introduced at a time when the Partido Revolutionario Institutional
(PRI) had already ruled for sixty years, and was just starting to experience its first real
challenges to its rule. Corruption was widespread, and rural areas were dominated by
local strongmen, the so called “Caciques.” In spite of having a decentralization law and
elected state and municipal councils and chief executives, Mexico remained a highly
centralized country in which over 90 percent of fiscal revenues were collected by the
state.
Under the DRD I project a Municipal Funds component10 was piloted in four Southern
States of Mexico, and quickly became its most successful components, eventually using
up 154 of the project loan’s 350 million dollars. As a consequence DRD II, implemented
from 1994 to 2002 was expanded to 8 states with a total of 1437 municipalities. The
Municipal Funds used up over 90 percent of the 500 Million dollars of the project loan.
The project introduced a large number of innovations in World Bank practices some of
which were specifically designed to counter the perceived weaknesses at the municipal
and community level:



The Municipal Funds program worked directly with rural municipalities as the
coordinators and implementers, and provided fungible grants to them.
A formula based on a municipal development index was used to allocate
money to the municipalities. This allowed targeting of the program toward the
poorest municipalities, and overcame the problems of political manipulation
associated (Diaz-Cayeros et al, 2004). In order to ensure that rural localities
receive a fair amount of money, the amount of money going to municipal
cabeceras was restricted, again using a formula.
A Municipal Development Council (MDC) was formed to coordinate and
supervise the program, and functioned under the guidance of the elected
municipal council, the Cabildo. The MDC was headed by the elected mayor,
and included the heads of localities within the municipalities, the
representatives of existing committees, and the Cabildo’s budget officer.
lack of technical quality in the preparation and appraisal of projects; and (iv) inadequate project
monitoring.
10
Municipal Funds is the generic term used to denote the DRD I & II component implemented initially as
part of Fondos Municipales de Solidaridad, which evolved in successive stages into: Fondo de Solidaridad
Munciipal (1995); Fondo de Desarrollo Social Municipal (1996-7); and, Fondo para la Infraestrictura
Social Municipal - FISM (since 1998 to present).
27



Projects emerged from community priorities and followed a clear project
cycle up to the approval at the MDC level. The MDC and the community then
signed a contract, and a technical implementation plan was prepared. The
municipalities then put the implementation responsibility and the money into
the hands of communities, i.e. using the empowerment model from the start.
New disbursement, procurement, and accountability procedures had to be
developed which were simple, yet provided safeguards for the proper use and
accounting of the resources transferred to the communities.11
Most community projects were small in size, and usually less than the
maximum of $50,000. Communities made their contributions in cash or in
kind, ranging from 20% to 40%, and fast completion of the works.
Due to its innovative nature, Municipal Funds activities used a learning-by-doing
approach across all levels of operation — federal, state, municipal and community levels.
Between 1995 and 1997 DRD II implemented around 106,000 projects in building or
rehabilitating classrooms and schools, potable water, rural roads, support to production,
electrification, sewage systems, pavement, urbanization, and others. In addition, Mexico
used its own funds to scale up the approach all over Mexico. Between 1990 and 2000
over half a million community projects were executed in this way. Poverty targeting to
municipalities was much better than under previous approaches. Projects were in line
with community priorities, generally of good quality, and about thirty percent cheaper
than projects implemented via state agencies.
Sustainability of the subprojects remained a problem, partly because the communities
were not yet legal entities capable of charging for services. Municipal performance varied
widely, depending on management capacity, which remains a major constraint for the
successful implementation of the now much enhanced municipal development budgets.
Participatory monitoring and evaluation remains weak and there exists no rigorous
impact evaluation of the program.
The relationship between the Mexican Government and the World Bank during the
design and execution of these programs were highly interactive, incorporated learning by
doing, as well as repeated disagreements and consensus building efforts. Mexico was not
dependent on World Bank lending to finance its poverty reduction program, but it used
the World Bank strategically to try new approaches and introduce many innovations into
its own programs. The program benefited from the existence of a legal framework for
Municipalities which needed activation, rather than creation. It also benefited from the
experience of many earlier rural development programs, and from the cadre of rural
development professionals these had formed and allowed to gain experience.
Despite the small size of the DRD program within Mexico’s overall poverty reduction
programs, many of the innovations introduced in the Municipal Fund approach were
11
Jean Claude Sallier and his Mexican colleagues invented and designed the original community-driven
disbursement, procurement, and financial accountability procedures without which the entire program
could never have been implemented.
28
incorporated into legislation supporting Ramo 33 and the corresponding Fiscal
Coordination Law of 1998 (for details see chapter….).
Brazil
In 1993 the World Bank supported rural development in ten North Eastern Brazilian
States via integrated rural development projects called “Programas de Apoio aos
Pequenos Produtores Rurais do Nordeste (PAPP).” These projects were supported by
loans to the Government of Brazil, which passed the resources to the Northeast
Development Agency (SUDENE), and from there to the states, which used their sector
agencies to implement the programs. There was no involvement of local governments.
The program was in crisis, suffering from many of the problems of area development
programs. Most components used either the service delivery or the intermediary model
for supporting rural communities, and by design about 55 percent of the project resources
were being used for the management of the projects at the federal, SUDENE, State, and
State agency levels. A significant proportion of the project resources were to be provided
as credit at close to market interest rates. Because Brazil was implementing the Plan Real
to overcome a major macro-economic crisis, real interest rates were at extremely high
real levels and therefore there was little demand for agricultural credit.
The projects, however, contained a successful pilot component, which provided matching
grants to small rural communities, the “Apoio às Pequenas Comunidades Rurais”
(APCR).12 These projects functioned very much like the community projects under the
Mexican Municipal Funds. In order to avoid the cancellation of the rural development
projects, it was decided to eliminate the credit component for which there was no
demand, reduce the number of intermediaries in the management of the project to just the
existing state technical units, and put all the project resources to support the APCR
approach, redesigned and adjusted to incorporate lessons from the Mexico Municipal
Funds.13 The intermediary model of community support was replaced by the
empowerment model. The redesigned projects became one of the largest and most
successful CDD programs in the World Bank at a total cost of 1.43 billion dollars, and
benefiting 37,592 communities comprising 2,540,733 benficiary families over the past 12
years. (Barboza et al 2006).
12
The component was designed by a team led by Hubertus von Pogrell, and incorporated lessons from
implementation experience in projects managed by Jean Claude Sallier, who invented the Community-base
disbursement and procurement procedures which made the Mexican DRD projects into a success.
13
The proposal for this very radical redesign was worked out in consultation with the locally recruited staff
of the Recife office of the World Bank by Hans Binswanger, at that time Senior Advisor for the Latin
America Region, Tia Duer, Division Chief for the Agricultural Division for Braizil, and Antonio
Maghalaes, undersecretary of Planning for Brazil. Luis Coirolo then led the team primarily from the Recife
office, which generated the buy in from the governors of the Northeastern States and the state level
technical units, produced the details of the redesign, including operational manuals and all other tools, and
brought it for approval to the Regional Management of the Latin America Region and the World Bank
Board. The buy-in from the Governors of the Northeastern States and Central Government officials was
secured by inviting them to visit the Municipal Funds Program in Mexico, which at that time was in its
third year of highly successful implementation.
29
Brazil had been under a military dictatorship until 1986, and had approved a new
democratic constitution in 1987 which introduced a radical program of decentralization to
States and Municipalities. Under the new constitutions, municipalities were to receive
twenty two percent of the public fiscal resources. The new dispensation also enabled
community and civil society participation in development. Nevertheless, at the time of
the redesign, municipalities and communities were still considered to be dominated by
small, self-serving elites, and most observers expected a high level of elite capture if they
were empowered with responsibility and fiscal resources.
Because of the fear of elite capture, the Brazil program took a more cautious approach to
the devolution of functions and funds to the municipalities. The state technical units still
functioned like social funds. However, the redesigned project introduced Municipal
Development Councils similar to those in Mexico in about 15 percent of the
municipalities, the so called FUMAC approach. These councils focused on identification,
sub-project approval and supervision, and did not manage money, which was directly
transferred from the State level to the communities. Attempts to devolve the funds to the
municipal level and have them allocated on a formula basis as in Mexico have so far been
resisted by the technical units of the states. As of this writing, the FUMAC model has
been generalized across almost all municipalities, and the Municipal Councils have
become involved in coordinating implementation of other federally financed programs.
The Northeast Rural development Program as a whole now has progressively been
improved by systematically addressing criticisms and concerns. Its enviable record of
success owes much to a high level of continuity in the Word Bank staff involved, and the
supervision and adaptation of the program from the Recife office of the World Bank. A
recent synthesis of evaluations conducted (Barboza et al 2006 a, b).
1. The loan resources actually reaching the communities have risen from around 45
percent to over 90 percent. At the same time the costs per sub-project have fallen
by around 30 percent. This means that for every dollar of total program costs,
about two and a half times as much project work can be implemented at the
community level than before the redesign.
2. The program has created significant social capital, at both the community and
local level.
3. As in Mexico, the program is well targeted to poor municipalities and to poor
communities, although it does not generally reach the poorest of the poor.
4. Project quality and sustainability is high.
5. The program has significantly impact in improving water supply and electricity in
poor communities of the Northeast, and has reduced child mortality and the
incidence of a number of diseases.
6. While rates of return to productive projects are high, they tend to have a higher
failure rate and lower sustainability than social and infrastructure projects.
7. The program has not had a significant impact on the net wealth of the
beneficiaries.
30
Brazil implemented what was to become called as Community-Driven Development
more cautiously than Mexico, and provided less fiscal resources to the municipal level. It
did not generalize the approach across the entire country by mainstreaming it into the
intergovernmental fiscal system. As a consequence the number of projects implemented
under the program remains much smaller than in Mexico. The challenge for Brazil and
the Bank is to take advantage of this experience by institutionalizing it more fully at all
levels: community, municipality, state, and the federation.
Indonesia
Steen: The Kacematan Development Program
Subsaharan Africa
Jacomina: Guinea, Burkina Faso, Uganda: The development of fully fungible local and
community development funds using norm-based allocations of resources, and directly
linking these funds to progress in decentralization.
2.3.5 The development of community-based disbursement and
procurement mechanisms
The progressive shift in large scale sectoral or multi-sectoral programs from intermediary
and service delivery approaches to an empowerment approach, in which communities
have full control over implementation and the finances associated with it, would not have
been possible using the classical disbursement and procurement mechanisms of the
World Bank and of other donors. As discussed before, Jean-Claude Sallier, together with
his Mexican counterparts, and supported by the procurement and disbursement specialists
of the Latin America Region of the World Bank, pioneered these systems. Six
innovations made this shift possible:
1. Legal ownership of the funds: Funds transferred to communities were considered
matching grants, and therefore became the property of the communities, rather
than the executing agency of the program or the World Bank. Just like credit
provided by a bank, the proper spending of these resources now becomes the
responsibility of the owner of the funds, the community
2. Replacing detailed accounting for the funds by a contract with the community:
The “expediente technico” for a small community project specifies what will be
done with the money, and how it will be used, as well as the technical details of
the sub-project to be financed. It is a four to six page document with a cover page,
which is in the form of a contract between the executing agency, the community,
and sometimes a facilitator or technical agent. At the end of the implementation,
the community signs a certificate of completion or handover of the project to the
community, in which the latter certifies that the project has been properly
executed and the funds accounted for. Rather than having to produce receipts for
each individual expense, it is this completion certificate which serves as the
“receipt” for accounting purposes of the executing agency and the World Bank.
Verification of the proper use of the funds by the implementing agency of the
31
3.
4.
5.
6.
program or by outside auditors then can consist of verifying that a road or a
classroom has actually been constructed and is in operation.
Direct transfer of the matching grants into the accounts of the community. This is
usually done in tranches, the first of which followed the signature of the contract,
and the second or third of which depend on demonstrating certain progress
benchmarks in the execution of the project.
Purchase technical support by the community. The community can select any
capable supplier and use a portion of the matching grant, usually of the order of 8
percent, to pay for the technical services.
Local shopping for both goods and services: This implies the suppression of the
traditional distinction between services and goods in the community procurement
rules. Local shopping is the main procurement system of communities for small
contracts and quantities of supplies. (Competitive bidding is still required if the
community entered into larger contracts where this method is justified). Local
shopping rules mean that community obtains offers from three suppliers, and that
its finance or management committee chooses from these three offers.
Transparency at the community level: Communities elect finance committees
which are in charge of day to day spending. Checks must be signed by at least two
members of the finance committee. The committee has to present all accounts to
the general assembly, which often also elects an audit committee to audit
accounts, purchases, stocks, and their uses.
Proper application of the rules means a radical shift in the direction of accountability in
project finance: From upwards accountability to the executing agency of the program and
the World Bank, to horizontal and downwards accountability to the members. Of course
upwards accountability remains, but in sharply simplified form. The idea of transferring
funds directly to communities met with widespread resistance at first, inside the Bank,
among borrowing countries, and among executing agencies and NGOs who were used to
intermediary and service delivery models of assisting communities. Therefore it took
over ten years form the initial development of these new procedures in Mexico to full
acceptance and widespread use in Social Funds, sector-specific and multi-sector CDD
programs it took nearly ten years. The final formulation of these rules in the World Bank
system occurred in ………. Jacomina, please find the date and the document which
sanctify these rules.
2.4 The future agenda for LCDD
From Messy Boutiques to Scaled up LCCD
Steen
Three strands evolved in parallel and led to much confusion within (three ways of
financing schools in Malawi) and across sectors (AG versus HD versus INF versus
Decentralization gurus) – Madagascar example. Little scaling up because of turf and
battles
32
What you need to discuss here is how the different approaches gradually incorporated
design elements from their rivals and therefore became more and more similar, most
strikingly as far as Social Funds, CDD, and Decentralization approaches are concerned.
Applying the guiding principles
Steen and Jacomina
Scaling up
Steen and Jacomina
References
Aiyar, Swaminathan, ed. August 1995. Decentralization: A New Strategy for Rural
Development. AGR Dissemination Notes. Number 1.
Aiyar, Swaminathan, ed. August 1995. How Well Has Decentralization Worked for
Rural Development. AGR Dissemination Notes. Number 2.
Aiyar, Swaminathan, ed. August 1995. Decentralization Can Work: Experience from
Colombia. AGR Dissemination Notes. Number 3.
Aiyar, Swaminathan, Keith McLean, and Suzanne Piriou-Sall, eds. September 1996. The
Political Economy of Democratic Decentralization. AGR Dissemination Notes.
Number 9.
Aiyar, Swaminathan, and Suzanne Piriou-Sall, eds. September 1996. How Rules And
Incentives Can Improve The Working Of Demand-Driven Rural Investment
Funds. AGR Dissemination Notes. Number 10.
Aiyar, Swaminathan, ed. September 1996. Has Decentralization Aided Biodiversity
Conservation? AGR Dissemination Notes. Number 11.
Banglapedia, 2005a?, Rural Development, Dhaka, Bangladesh
Banglapedia, 2005 b? The Comilla Model, Dhaka, Bangladesh
Barboza, Tulio, Hans Binswanger, Alberto Costa, and Claudia Romano 2006: Hans to
provide title
Campfens, Hubert, ed., Community Development Around the World, University of
Toronto Press, Toronto
Diaz Cayeros, Alberto and Beatrice Magaloni, 2004, The Politics of public spending, part
II, Background paper for the World Development Report, World Bank, Washington DC
Dongier, Philippe, Julie van Domelen, Elinor Ostrom, Andrea Ryan, Wendy Wakeman,
Anthony Bebbington, Sabine Alkire, Talib Esmail, and Margret Polsky, 2003,
33
“Community Driven Development,” Chapter 9 of PRSP Sourcebook, Washington
DC
Hegde, N.G. 2000, Community Development in India : An Overview, Proc. of the Asian
Productivity Organization, Tokyo, Japan seminar on Comparative Study on Planning
Process of Community Development, Ulaanbaatar, Mongolia. Aug. 21-26
Khwaja, A.I. 2000. “Can Good projects Succeed in Bad Communities?” Harvard
University. Paper Presented at the Annual Bank Conference on Development Economics.
May 2
Krishna Anirudh 2004. Partnership Between Elected Local Governments and
Community-Based Organizations: Exploring the Scope for Synergy. Social Development
Papers Community Driven Development Number 52/February.
Kumar, N. (2002), The challenges of community participation in forest development in
Nepal, OED Working Paper Series, World Bank, Washington D.C.
Lin, Justin and Funning Zhong. 1997. Fiscal Decentralization and Rural Development
in China. Forthcoming.
Manor, J. 1999. The Political Economy of Democratic Decentralization. Directions in
Development. Washington D.C., The World Bank.
Mansuri, G. and V. Rao (2004), ‘Community-Based and Driven Development: a critical
review’, The World Bank Research Observer 19 (1): 1-39.
McLean, Keith, Graham Kerr, Suzanne Piriou-Sall and Melissa Williams, 1998
Decentralization and Rural Development: Characterizing Efforts In 19 Countries,
Washington DC, World Bank working paper
Narayan Deepa 1995. The Contribution of People’s Participation: Evidence from 121
Rural Water Supply Projects. Environmentally Sustainable Development Occasional
Paper Series No.1. The World Bank.
Parker, Andrew N. 1995. Decentralization: The Way Forward for Rural Development?
Policy Research Working Paper Series, No. 1475. Washington, D.C.: Policy
Research Dissemination Center, The World Bank.
Piriou-Sall, Suzanne. December 1996. Demand-Driven Rural Investment Fund: A New
Generation of Social Funds. AGR Dissemination Notes. Number 13.
Platteau J-P. and F. Gaspart. 2003. ‘The Risk of Resource Misappropriation in
Community-Driven Development’, World Development 31 (10): 168-1703.
34
Platteau Jean-Philippe. 2003. Institutional and Distributional Aspects of Sustainability in
Community-Driven Development. Fifth Biennial World Bank Conference on Evaluation
and Development “Evaluating Development Effectiveness: Challenges and the way
Forward. Washington DC
Romano, Claudia, Alberto Costa, Naercio......, and Hans Binswanger, 2006, Impact
evaluation: Hans to provide title and details.
Staples, Eugene S., 1992, Forty years: A Learning Curve: The Ford Foundation in India,
1952-1992, Ford Foundation, New York
UNCDF, 2005, Delivering the Goods: Building Local Government Capacity to Achieve
the Millennium Development Goals, A Practitioner's Guide from UNCDF
Experience in Least Developed Countries. New York
World Bank. 2000. The Community Driven Development Approach in the Africa region:
A Vision of Poverty Reduction Through Empowerment. Washington, D.C.: World Bank.
World Bank, 2001a, Attacking Poverty, World Development Report, Washington DC
World Bank 2001b. Empowering the Poor through Decentralization: Brazil Rural Poverty
Alleviation Program.” Social Development Note No. 51. March 2001.
World Bank, 2001c, Promoting Good Governance through Social Funds and
Decentralization, Prem Notes # 51, Washington DC.
World Bank 2001d. Decentralization and Governance: Does decentralization improve
public service delivery? Notes Number 55.
World Bank 2001e, Sourcebook for Community-Driven Development in Africa,
Washington DC
World Bank 2002a, Empowerment and Poverty Reduction: A Sourcebook, Washington,
D.C.: World Bank
World Bank 2002b. Social Funds: Assessing Effectiveness. Operations Evaluation
Department.
World Bank, 2003a, Making Services work for the Poor, World Development Report,
Washington DC
World Bank, 2003b, The World Bank Participation Sourcebook. Environmentally
Sustainable Development. http://www.worldbank.org/wbi/sourcebook/sbhome.htm, First
published in 1996, Washington DC.
35
World Bank 2003c. Community-Driven Reconstruction as an Instrument in War-toPeace Transitions. CPR Working Papers. Paper No. 7 August. Social Development
Department ESSD Network.
World Bank 2004. Community-Driven Development in Urban Upgrading. Social
Development Notes. February, No.85
World Bank 2004a. Community-Driven Development in Local Government Capacity
Building Projects: Emerging Approaches in Africa. Social Development Notes November
86
World Bank 2004b. Reducing Poverty Sustaining Growth. Scaling Up Poverty
Reduction. A Global Learning Process and Conference in Shanghai May 25-27 2004
World Bank, 2005a, Community-Driven Development, http://www.worldbank.org/CDD
World Bank, 2005b, Linking community empowerment, decentralized governance, and
public service provision through a local development framework, Washington DC
World Bank, 2005c, The Effectiveness of World Bank Support for Community-Based
and -Driven Development, Independent Evaluation Group, Washington DC
World Bank, 2005d, Community-driven development website. First pubished in 2001?
(Steen to provide date when first put up) Washington DC
World Bank, 2005e, The Online Sourcebook on Decentralization
(http://www.ciesin.org/decentralization/Entryway/english_contents.html). First published
in 199? (This was maintained by Melissa Williams, and we need to find out when it was
started)
36
37