Challenges to Decentralisation of Agricultural Extension
Site Index
Main Page
Issues in Program Design

Contributor: Food and Agriculture Organization of the United Nations
Contact: Jean Bonnal

Challenges to Decentralisation of Agricultural Extension

Issues And Challenges

Decentralised extension services has long been a feature of federal countries. Many developing countries are now decentralising extension in the expectation that the service will be closer to the client, and thus more relevant. Budgetary constraints also play a role in the decision. There are three main reasons why governments decentralise agricultural services (Smith, 1997):

  • a desire (or demand) to roll back the role of the State due to central government failings or complexity of local issues;
  • an inability of the State to continue to finance a whole range of services, and;
  • a view that democracy is best served through devolved functions with enhanced participation at local level.

If those are the reasons for decentralising agricultural services, disappointment is likely. A weak state will not provide the basis for effective decentralisation. Strong central institutions are needed to create a framework within which it is easier for local government to operate, and a shortage of administration expertise at central level is bound to be even worse at local level. Similarly the financing constraint is likely to be even greater for local governments, who find it difficult to raise taxes to pay for local services or to impose sanctions on those unwilling to pay. In practice, fiscal decentralisation may provide central governments with a convenient excuse for abandoning certain functions and does not guarantee improved delivery.

In the case of extension, however, decentralisation does seem to offer particular hope for improving relevance and responsiveness. These, together with sustainability, should be the main criteria to rate the success of decentralised extension. Many problems and solutions are location specific. In terms of relevance this should give a clear advantage to the local provision of advice. However, administrative boundaries rarely coincide with agro-ecological zones (nor with socio-economic situations): there may be a large diversity of situations within the purview of a local government, while the capacity to adjust the advice to local conditions (or to specific groups) may be negatively affected by decentralisation. In particular, good linkages with agricultural research may be difficult to establish at local level if there is no research facility in the region. Tunisia offers an example of a largely decentralised extension system in the context of a centralised research environment. The new Regional Research Poles may remedy this, if they receive the resources and authority to conduct work based on regional needs.

Similarly, responsiveness to farmer problems may not automatically result from decentralisation. Extension managers become closer to the client but not necessarily more attentive to their problems. Staff attitudes need to change, and farmers need to get organised to make them heard. In the Philippines, some farmers were able to get advantage of tailored made extension activities. But for most upland farmers, there has been no change in a system that never met their needs adequately(Malvicini, 1996).

The other areas where centralised extension is generally weak, such as limited coverage, low effectiveness and low cost efficiency, are less likely to be improved directly by decentralisation. However, decentralised services in Colombia increased coverage by a factor three and cost-efficiency also improved. This is a rare case where decentralisation was accompanied by increased resources: the number of staff was multiplied by four (Garfield et al., 1997).Rivera (1996) sees three main policy directions dominating the decentralisation of extension: structural reform to improve institutional responsiveness and accountability, fiscal decentralisation to share extension costs with local governments, and farmer participatory involvement in decision making and management of extension. These policy objectives echo the three motives indicated earlier for decentralising, and suggest possible remedies to the risks they entail. But will governments who want to reduce extension budgets be willing to initially maintain the same levels of spending and delegate control over it?

The financial issue is central, and compounded by the fact that decentralisation is occurring in a context of decreased faith in (public) extension. In this context, the central level is unlikely to provide the support required to facilitate decentralisation. At the local level, where extension needs to be re-created as a farmer-driven service, farmers have also learned to be sceptical. Will they invest in it and get organised to obtain the best out of it? Will they accept to share costs, as needed to compensate for the decrease in public funding, and also to ensure accountability? And if so, can cost-sharing be done without prejudice to poor farmers or to other key policy objectives (e.g. environmental protection)?

It is already apparent that decentralisation of extension is unlikely to fulfil the excessive expectations it has aroused. But the outcome is hard to predict, because the circumstances and the modalities of decentralising extension services vary so much. Hence the need to document and analyse existing experience. This study remains to be done. We only wish her to point to some areas which need particular attention, and highlight a few hypotheses which require testing.

Areas of Concern and Information Needs

In many countries, decentralisation is ongoing, or just starting. It will take some years before the dust settles and an objective and well-informed picture emerges. Information is often anecdotal and does not conform to a common framework which would facilitate comparisons and give confidence in drawing lessons. The context of decentralisation is rarely documented. If government is already largely decentralised, the reform of extension is likely easier as institutions and mechanisms are already in place. Also, extension does not operate in a vacuum. We need to know how the relationships with other services are taken into account when extension is decentralised. Finally, analysts rarely distinguish between transition problems and structural issues, i.e. those which pre-existed decentralisation but may become more apparent after. The distinction is admittedly difficult when the process is still ongoing; and may be somewhat artificial as the former could become structural if not addressed early enough. Transitory difficulties distract attention from the deeper issues, which is a pity since decentralisation may be a unique opportunity to tackle those.

Transition Difficulties

Because decentralisation of extension is often part of a sweeping reorganisation of government central services, the transition problems are not specific to extension. Essentially, they relate to the allocation and channelling of funds to local governments, and to staff management. Devising and implementing new allocation procedures is complex and rarely a smooth process, as expectations relating to the use of funds are bound to differ between local and central levels (otherwise, why decentralise?). For the staff devolved to lower government levels, a simple hierarchical structure and uniform status gives way to a complex line of command, typically involving both the line ministry and the local government, and to regional differentiation of status.

Uganda illustrates those transition difficulties. District councils are given primary responsibility for allocating budget resources for extension. But the budget mechanism, flow of funds to district level and financial management procedures still need to be clarified, five years in the process. Moreover, resources available are just enough for salaries, leaving little for staff development and extension activities. The staff are now employed by the district councils, but technically under the ministry of agriculture, who has delegated its responsibility to the National Agricultural Research Organisation (NARO). NARO is not prepared to the new task, and may see the extension service only as an arm for transferring research results to the field. This would run counter to the intention of decentralised, farmer responsive services. The role of NARO should rather be one of supporting staff engaged in locally programmed activities.

In the Philippines, a recent survey (Malvicini, 1996) shows that devolved extension staff have lost career development perspectives; salary increases are blocked to progressively align remuneration with that of other municipal employees who have lower education levels; and there are no funds for staff training. Another issue is the politicisation of new recruitment. As a result, quality of services is decreased.

Structural Issues

The main areas of concern relate to the structural issues of accountability, relevance, the lack of a constituency, and quality.

  • Accountability. Decentralisation does not give farmers control over the management of extension or over the appointment of individual staff. Specific mechanisms are needed to ensure participatory management of programmes by farmers. In Colombia, the law that created the UMATAs (municipal extension offices) also required the creation of farmer dominated Municipal Committees for Rural Development to co-ordinate and prioritise rural development activities, including extension. However, less than half the municipalities have created these committees, and they seem mostly powerless.

A related issue is that farmers are not an homogenous lot; decentralised, public extension cannot address the needs of all groups. The role of the public sector must be redefined to permit multiple approaches which account for user diversity, and to develop partnerships with farmer organisations, NGOs and the private sector for service delivery (World Bank, Participation Sourcebook, Washington 1995).

China has raised accountability through various cost-sharing mechanisms. The most radical approach features contractual sharing of benefits and risks between farmers and an extension agent on a project they have negotiated together. This is applied mostly in horticulture, cash crops and livestock production (Yonggong 1998).

  • Relevance. Technical relevance is not necessarily improved by decentralisation. Van Crowder (1996) already raised the question of what effect decentralisation has on the research-extension linkages: does it result in a de-coupling of research-extension linkages, or does it actually improve linkages. Technical relevance requires a local research capacity to verify and adapt technologies to farmer conditions. NARO in Uganda is setting up a network of Outreach Centres. It is not yet clear if these centres will be able to move beyond their present role of technology dissemination towards supporting a farmer-led, problem-solving approach. Now, it is fair to recognise that research-extension linkages rarely worked well in a central setting. Decentralisation may be an opportunity to revisit the research-extension nexus in the light of more global concepts such as the Agriculture Knowledge and Information System (AKIS), encompassing research, education and training, as well as traditional knowledge, and featuring multiple communication and delivery channels.
  • Constituency. Agricultural extension is not a high priority for local governments. Health, schools and public infrastructure are seen more important. In Uganda, a 1997 survey showed that local governments did not allocate resources as expected to the (centrally defined) priority programs. Only three out of 32 districts surveyed had budgeted over 3% of their expenditure on agriculture; most had budgeted to spend only one percent (Uganda 1998). Government response has been to tie financial transfers through conditional grants, which runs counter to the o bjective of decentralisation. The issue is not simply one of farmer representation in local assemblies. The central extension services were rarely seen by the majority of farmers as meant for them. To build a constituency among farmers requires important changes from the staff. They must change attitudes and start listening to farmers. They must also acquire the skills (diagnostic, participatory technology development, etc.) needed for identifying and for meeting farmer needs.

Building a constituency for decentralised extension is a condition of financial sustainability. In Colombia, where costs to the central government have initially increased, the aim is to gradually transfer the fiscal burden to municipalities, once the demand for extension services has been created. Meanwhile, municipalities are obligated by law to have an extension office.

  • Quality. Concern of degrading quality is a recurrent theme. In Colombia, this is attributed to looser links with research, political recruitment, and reduced use of private consultants. The reason for the latter is again political: directly employed staff are more easily controlled. However, one aspect of quality has improved: the extension agents are now seen closer to small farmers. They are in every municipality and part of the local government, which is more accessible to farmers than the central government.

In the Philippines, the quality concern is linked to the cessation of training, the lack of career development opportunities and of incentives for improvement, and again the politicisation of staff recruitment. Training is needed to facilitate the re-orientation to community-based approaches; if not provided, staff may become "obsolete" and extension may collapse at the local level (Malvicini 1996). Malvicini argues that a strong central level is crucial to quality, as it can provide and enforce suitable norms regarding staff recruitment, management and training, facilitate linkages with research and the free flow of knowledge among regions and outside, and develop effective quality control and monitoring systems. But surely, this centralised vision of quality control is mostly relevant to the transition stage, until organised beneficiaries can have a large say in what quality means for them.

Hypotheses That Require Testing

  • A general hypothesis is that, unless the structural issues of extension are addressed, there are only limited benefits from decentralisation - essentially in the form of shorter decision lines and greater proximity to the client - and these may be more than offset by politicisation, reduced funding, disruption of knowledge linkages, etc. which all affect quality.
  • Decentralised extension is unlikely to reduce costs in the short term. Increased resources are needed to re-orient staff and provide incentives in a reform which is feared by many; and to energise other support services without which extension staff remain powerless to meet farmer expectations. Only if the new extension is appreciated is there a possibility that local funding (including user contributions) may in the future be willingly spared for the service;
  • Decentralisation of extension is possible, and more likely to succeed, in the absence of intergovernmental decentralisation. Although it often takes place in the context of a general reform of the state, this need not be. If extension is restructured separately, it may be easier to take into account its specificity and address the crucial issues.

The reorganisation of extension services in Zaire (now Democratic Republic of Congo) in the early 1990's is an example of such a decentralisation in a very difficult context. After restructuring, the system included four types of organisations (FAO 1997):

  • the "Service National de Vulgarisation" (SNV), with a National Directorate and 11 Regional Co-ordinations, whose role is in planning, co-ordination, training, methods and control;
  • a network of public and private providers. In 1995, 121 different organisations participated in extension: the public sector, with 22 organisations in 11 regions represented 57% of total extension staff; parastatals (48 agencies and projects), contributed 24%; and the private sector (51 NGOs), 18%;
  • service organisations with which the SNV established partnerships to ensure essential support services in research, seed production, marketing, input procurement, etc. Many of these were producer organisations, which the SNV helped technically and financially by mobilising external aid resources through mini-projects;
  • farmer groups, working with the extension specialist on a community project to define an annual programme, carry out demonstrations and other extension activities, and to prepare mini-projects for solving problems related to the adoption of new technology.

Such a form of decentralisation provides mechanisms that improve accountability, relevance and cost-efficiency.

  • Decentralised extension requires decentralised and participatory research. Just re-arranging disrupted research-extension linkages will not provide decentralised staff with the support they need for solving farmers problems. Research must focus on farmer needs, farmers need to be involved in planning implementing and evaluating research activities.

Indonesia is experimenting with a new approach to decentralised adaptive research through Agricultural Technology Assessment Institutes (BPTP) integrating researchers and extension specialists under one roof to assess new technology under farmer conditions and develop solutions to farmer problems. This is breaking the tradition of a top-down, linear research-extension-farmer relationship and has potential to develop instead a mutual working pattern where the three groups act like a team with a common objective, if and when the reform is completed and the BPTPs become financially autonomous from line agencies.

Under a CIAT initiative, Bolivia, Ecuador, Honduras and Peru have seen the development of locally elected committees (CIALs), composed of experimenting farmers who manage and conduct research on behalf of the community as a whole. In Colombia, where the CIALs started in 1990, a number of these have evolved into local seed enterprises. Integration of research, extension and input supply services is thus achieved at community level (Ashby, 1995).

  • In the complex local political landscape created by decentralisation, where local governments need to assert their authority vis--vis traditional structures and/or new grass-root development organisations (e.g. NGOs, producer groups, etc.), extension is probably less at stake than services involved in natural resources protection. However, controlling extension workers may be seen as conferring an advantage to local politicians. Where this is the case, devolving responsibility to user groups might be strongly resisted by local governments.


Ashby, J.A., 1995. Institutionalising Farmer Participation in Adaptive Technology Testing with the "CIAL". Agricultural Research and Extension Network Paper No 57, ODI, London.

Crook, R. and Manor, J., 1994. Enhancing participation and Institutional Performance: Democratic Decentralisation in South Asia and West Africa. ESCOR, ODA.

FAO, 1997. Vers une vulgarisation participative. Etude de cas: le Zare. Rome, 1997.

Garfield, E., Guadagni, M. and Moreau, D., 1997. Colombia: Decentralisation of Agricultural Extension Services, World Bank Extension Website

Malvicini, P. et. al., 1996. Decentralisation of Agricultural Extension in the Philippines: Forming Community-Based Partnerships. Cornell University, Ithaca N.Y.

Rivera, M.R., 1996. Agricultural Extension in Transition World-wide: Structural, Financial and Managerial Strategies for Improving Agricultural Extension. Public Administration and Development, Vol. 16, 151-161.

Smith, L D., 1997. Decentralisation and Rural Development: The Role of the Public and Private Sector in the Provision of Agricultural Services. Technical Consultation on Decentralisation, FAO, Rome, 16-18 Dec. 1997.

Uganda (Government of), 1998. Towards a Sector Wide Approach: Developing a Framework for the Modernisation of Agriculture in Uganda. Statement to the December 1998 Consultative Group meeting.

Van Crowder, L., 1996. Decentralised Extension: Effects and Opportunities, FAO Sustainable Development Site -

Yonggong, L., 1998. Institutional and Policy Reform of Rural Extension in China During the Transition Towards a Market Economy. Training for agricultural and rural development 1997-98. FAO, Rome.